KR100530853B1 - Device for encoding video information of mpeg-4 encoder and thereof method - Google Patents

Abstract

The present invention provides an image information encoding apparatus and method thereof, and in particular, an image information encoding apparatus of an MPEG-4 encoder to which an object boundary block merging / separation technique using spatial surplus of original image information and image error information existing in and between blocks is applied. And an image information encoding apparatus of an MPEG-4 encoder according to the present invention, and a method for reconstructing shape information whether or not to merge or separate an object boundary block when encoding region-based image information. By using the object boundary block merging / separation information extracting unit to separate the object boundary block merging / separation information from the object boundary block merging unit and the merged block separation unit, the redundant configuration can be avoided. This has the effect of reducing the complexity of the time.

Description

Image information encoding apparatus of MP4 encoder and its method {DEVICE FOR ENCODING VIDEO INFORMATION OF MPEG-4 ENCODER AND THEREOF METHOD}

The present invention relates to an image information encoding apparatus and a method thereof, and more particularly, to an image of an MPEG-4 encoder to which an object boundary block merging / separation technique using spatial surplus of original image information and image error information existing in and between blocks is applied. An information encoding apparatus and a method thereof are provided.

MPEG Group, which recently developed and decided on MPEG-1 and MPEG-2, the international standard for image and audio encoding technology and system configuration, has decided on MPEG-, the international standard for the next generation image and audio encoding technology and system configuration. 4 is under research and development.

The development of MPEG-4 began with the need to support the next generation of video and audio applications that could not be supported by existing standards, and new methods for video and audio data communication, access and manipulation (e.g. Object-oriented interactive functions and connections through other networks), and provide useful characteristics in communication environments where errors are easily generated and communication environments with low data rates.

In addition, the MPEG-4 technology integrates computer graphics technology and provides functions for encoding and manipulating natural / artificial images and sounds together, and low cost, newly developed or expected to be developed due to rapid expansion and improvement of multimedia information. It has an extensible and open structure to support the functions required for high-performance applications, among which is the IMPROVED COMPRESSION EFFICIENCY.

Applications currently expected to apply MPEG-4 technology include Internet multimedia (IMM; INTERNET MULTIMEDIA), interactive video games (IVG; INTERACTIVE VIDEO GAMES), video conferencing and video telephony (IPC; INTERPERSONAL COMMUNICATIONS). ), INTERACTIVE STORAGE MEDIA (ISM), MULTIMEDIA MAILING (MMM), WIRELESS MULTIMEDIA (WMM), NETWORKED DATABASE SERVICE (NDB), ATM (RES; REMOTE EMERGENCY SYSTEM) and REMOTE VIDEO SURVEILLANCE (RVS).

In addition, MPEG-4 technology requires that users can communicate only with the desired object in the video, access to search and read, and cut and paste to support existing or future applications. Should support this.

1 is a hardware block diagram of a typical MPEG-4 video encoder, and has a structure different from that of the H.261, MPEG-1, H.263, and MPEG-2 video encoders. The general MPEG-4 will be described in detail with reference to FIG. 1.

VOP (VIDEO OBJECT PLANE; hereinafter referred to as VOP) forming unit 100 serves to form a VOP, meaning a content object of any shape that can be accessed and edited by the user at a point in time on the time axis, and moves The estimator 200 receives image information of the VOP formed by the VOP forming unit 100, receives image information of the previous VOP reproduced by the VOP detector 800, and reconstructs the shape of the shape information encoder 300. It receives the information or the circular shape information and outputs the motion information estimated motion in macroblock units.

The motion compensator 400 receives the motion information estimated by the motion estimator 200, receives the shape information reconstructed by the shape information encoder 300, and reproduces the previous VOP reproduced by the VOP detector 800. Image information of the motion-compensated VOP is received by receiving the image information, and the subtractor 500 receives the image information of the motion-compensated VOP from the motion compensator 400 and at the same time the VOP forming part 100. It receives the image information of the formed VOP and plays a role of calculating the difference between the two signals.

The image information encoder 600 receives the difference value from the subtractor 500 to encode the macroblock in 8 × 8 block units, and the adder 700 compensates for the motion in the motion compensator 400. It receives and adds the image information of the VOP and the internal information of the object encoded by the image information encoder 600.

The VOP detector 800 detects the image information of the previous VOP by receiving the addition value from the adder 700, and the detected image information of the previous VOP is the motion estimator 200 and the motion compensator ( 400 is used for motion estimation and motion compensation.

The shape information encoder 300 receives the shape information of the VOP from the VOP forming unit 100, encodes, decodes, and outputs the original shape information.

The coded shape information is input to the multiplexer 900 and multiplexed with the encoded image information, and the decoded shape information (restored shape information) and the original shape information are displayed as a dotted line. The motion compensation unit 400 may be input to the motion compensator 400 and the image information encoder 600 to encode motion estimation, motion compensation, and internal image information of an object.

The multiplexer 900 multiplexes the motion information estimated by the motion estimation unit 200, the image information of the object encoded by the image information encoder 600, and the shape information encoded by the shape information encoder 300. The buffer 910 serves to transmit the multiplexed signal in the multiplexer 900 as a bitstream.

The biggest feature of the MPEG-4 video encoder is that the introduction of the shape information encoder 300 and the concept of VOP is different from the existing world information encoding standard.

FIG. 2 is a diagram illustrating a VOP, which is an image composed only of information in a whole image or a meaningful object region constituting the image in an arbitrary image, and reconstructed into macroblocks composed of 16 × 16 pixels as shown in FIG. 3 for image encoding. do.

FIG. 3 is a diagram illustrating the types of macroblocks in a VOP. When the VOP is reconstructed into macroblocks, three types of macroblocks exist. The macroblock in the object composed of only the information in the object and the information in the object are not included. Out-of-object macroblocks and object-bound macroblocks that partially contain information within the object.

FIG. 4 is a diagram showing the types of blocks in a VOP, in which there is an intra-object block composed of only the information in the object, an out-of-object block having no information in the object, and an object boundary block partially having the information in the object.

The macroblock in the object is composed of only the blocks within the object, the macroblock outside the object is composed of only the blocks outside the object, the object boundary macroblock may be three kinds of blocks consisting of 8 × 8 pixels. , Block inside object, block outside object, object boundary block.

On the other hand, in order to encode the original image information and the image error information of the object boundary block, the pixel values of the pixels in the region outside the object must be determined first, and the image error signal (RESIDUAL TEXTURE DATA) and the original image information ( INTRA TEXTURE DATA) is called texture information.

Arbitrary shape VOP texture information encoding techniques considered until recently include padding techniques using padding and techniques that do not use padding. As padding techniques, low frequency extension (LPE; LOW PASS EXTRAPOLATION) applied when encoding image information in an image Technology and ZERO PADDING technology that fills zero value when encoding image information between images, and coding technology without padding uses shape-adaptive DCT (PID) that uses only intra-object information without padding. SHAPE ADAPTIVE DISCRETE COSINE TRANSFORM) technology.

FIG. 5 is a view for explaining a low frequency component extended padding technique in an area-based image encoding method, and receives image error information or original image information (a) from the subtractor 500 of FIG. An average value calculator 612 which receives the padding shape information (eg, original image shape information, reconstructed shape information, or newly formed shape information using both information) from 300 to obtain an average value of only signals in the object; The average value padding unit 613 for padding the outside of the object area with the average value obtained by the average value calculator 612, and the output signal (b) of the average value padding unit 613 is input only to the pixels belonging to the padded area outside the object. The result is obtained by adding the pixel values in the 8 × 8 block among the pixels in the top, bottom, left and right, and dividing the pixel value by the number divided by the number of pixels in the 8 × 8 block among the pixels in the top, bottom, left and right. Outputting The low frequency filtering unit 614 is configured.

FIG. 6 is a diagram for describing a zero padding technique in an area-based image encoding method, and receives image error information or original image information from the subtractor 500 of FIG. 1, and has a padding shape from the shape information encoder 300. It consists of a zero value padding unit 615 that receives information (eg, original image shape information, restored shape information or newly formed shape information using two pieces of information) and pads an area outside the object to zero value. have.

FIG. 7 is a view for explaining a shape-adaptive DCT technique in an area-based image encoding method, and upwards as shown in (b) for intra-object image information (original image information or image error signal) as shown in (a). After moving, the vertical one-dimensional DCT is performed, and the coefficient formed as shown in (c) is moved to the left as shown in (d), and the horizontal one-dimensional DCT is performed as shown in (e) as shown in (f). Shape-adaptive DCT coefficients are formed.

The DCT coefficient after the shape-adaptive DCT is quantized and a zigzag scan is applied. The DCT coefficient is a technique of encoding by performing an adaptive scan that skips an area where no coefficient exists. Unlike padding technology, shape-adaptive DCT encodes using only intra-object image information in the object boundary block.

In addition to the padding technique and the shape-adaptive DCT technique, an object boundary block merge / separation technique is applied to increase the VOP coding efficiency of an arbitrary shape. Currently, the image information encoder may be configured as shown in FIG. 8 or 9 by the object boundary block merging / separation technology, and has been adopted as a technology for MPEG-4 image coding version 2 world standardization.

The object boundary block merging / separation technique includes an object boundary block merging process that can improve encoding efficiency by using spatial surplus of intra- and inter-image information existing between 8 × 8 blocks as well as adjacent 8 × 8 blocks. It is a video information encoding method using a merged block separation process as a reverse process of the merge.

8 is a block diagram of an image information encoder according to an object boundary block merging / separation technique to which a conventional padding technique is applied. The out-of-object region padding unit 610, the object boundary block merging unit 620, the DCT unit 630, and quantization are illustrated in FIG. A unit 640, a DCT coefficient and related information encoder 650, an inverse quantization unit 660, an inverse DCT unit 670, and a merged block separation unit 680.

The out-of-object region padding unit 610 receives original VOP image information or image error information from the subtractor 500 of FIG. 1, and shape information (original image shape information, reconstruction) for padding from the shape information encoder 300. The shape information or the original image shape information and the arbitrary shape information generated by using the restored shape information).

The object boundary block merging unit 620 receives image information from the out-of-object region padding unit 610 and shape information restored from the shape information encoding unit 300, and determines whether to merge with the shape information. It merges the object boundary blocks of the image information when merging is possible.

The DCT unit 630 receives the image information obtained by merging the object boundary blocks from the object boundary block merging unit 620 to perform DCT, and the quantization unit 640 performs the DCT in the DCT unit 630. DCT coefficient is input and quantizes.

The DCT coefficient and the related information encoder 650 receive the variable quantized DCT coefficients by the quantizer 640 and encode the related information while outputting the encoded information to the multiplexer 900.

The inverse quantization unit 660 receives inverse quantization by receiving the quantized DCT coefficients from the quantization unit 640 and outputs the inverse quantized DCT coefficients.

The inverse DCT unit 670 receives inverse quantized DCT coefficients from the inverse quantizer 660 and performs inverse DCT conversion to output a decoded signal.

The merged block separator 680 receives the decoded image information from the inverse DCT unit 670, receives the shape information restored by the shape information encoder 300, and determines whether or not to separate the received image information. Separating the merged block serves to output to the VOP detector 800 through the adder (700).

9 is a block diagram of an image information encoder according to a conventional object boundary block merging / separation technique to which a shape adaptive DCT is applied. The object boundary block merging unit 611, the shape adaptive DCT unit 621, and the quantization unit 631 are shown. ), A shape adaptive DCT coefficient and related information encoding unit 641, an inverse quantization unit 651, an inverse shape adaptive DCT unit 661, and a merged block separation unit 671.

The object boundary block merging unit 611 receives the original image information or the image error information from the subtractor 500, receives the restored shape information of the object boundary block, and determines whether or not the object boundary is merged based on the shape information. It merges the object boundary blocks of visual image information.

The shape-adaptive DCT unit 621 receives reconstructed shape information defining an object area from the shape information encoder 300 and receives image information obtained by merging an object boundary block from the object boundary block merging unit 611. The shape adaptation DCT is received and the quantization unit 631 receives the shape adaptation DCT coefficient from the shape adaptation DCT unit 621 and quantizes the shape adaptation DCT.

The shape-adaptive DCT coefficient and the related information encoder 641 skips the quantized shape-adaptive DCT coefficient in the region where no coefficient exists and applies a zigzag scan to the quantization unit 631 (ADAPTIVE). SCAN) to perform encoding, to encode related information, and to output signal information to the multiplexer 900.

The inverse quantization unit 651 receives the quantized signal from the quantization unit 631 and performs inverse quantization. The inverse shape-adaptive DCT unit 661 is inverse quantized by the inverse quantization unit 651. The shape information encoder 300 receives the restored shape information defining the object region from the shape information encoder 300 and performs inverse shape adaptation DCT.

The merged block separator 671 receives the decoded signal from the inversely adaptive DCT unit 661, and uses the shape information used by the object boundary block merging unit 611 from the shape information encoder 300. It receives the restored shape information for merging / separation of the object boundary block, which is the same shape information, and separates the merged block and outputs it to the VOP detection unit 800 through the adder 700.

However, the image information encoder according to the object boundary block merge / separation technique using the conventional padding technique or the shape-adaptive DCT receives the shape information restored from the shape information encoder in the merge process and the separate process, respectively. After the determination, the merging and separating of the object boundary blocks of the image information is performed, which causes a complicated configuration when implementing the system.

Accordingly, an object of the present invention is to solve the above problems, and an object of the present invention is to provide an image information encoding apparatus and method for an MPEG-4 encoder, which can avoid a redundant configuration and reduce complexity in system implementation. To provide.

In order to achieve the above object, the video information encoding apparatus of the MPEG-4 encoder according to the first embodiment of the present invention receives the original image information or the image error information and uses the shape information for padding to remove the object from the image information. The object boundary that extracts the object boundary block merging / separation information according to whether to merge the object boundary block and whether to separate the object boundary block by receiving the rest area padding unit and the restored shape information. Object boundary using a block merge / separation information extractor, shape information restored and received from the object padding region outside the object padding unit, and merge / separation information input from the object boundary block merge / separation information extractor An object boundary block merging unit for merging image information of blocks, a DCT unit for receiving DC information from the object boundary block merging unit and performing DCT; A DCT coefficient that is variably encoded using a merger / separation information input from the object boundary block merger / separation information extraction unit by receiving a quantization unit that receives and quantizes a DCT coefficient from the DCT unit. And a related information encoder.

The video information encoding method of the MPEG-4 encoder according to the first embodiment of the present invention for achieving the above object is to pad the area outside the object using the shape information for padding the original video information or video error information. A first step and a second step of merging the image information of the object boundary block using the merged / separated information extracted from the image information padded in the first step and the restored shape information; and the image merged in the second step. And a fourth step of performing a DCT on the information and quantizing the information, and a fourth step of variable length encoding the DCT coefficients quantized in the third step by using the merged / separated information extracted.

In order to achieve the above object, the image information encoding apparatus of the MPEG-4 encoder according to the second embodiment of the present invention receives the restored shape information and determines whether or not to merge the object boundary block. Object boundary block merging / separation information extractor for determining and extracting object boundary block merging / separation information, and shape information restored by receiving original image information or image error information and object boundary block merging / separation information extracting unit The object boundary block merging unit for merging the image information of the object boundary block using the merge / separation information input from the object and the object boundary block merging / separation information extracting unit which receives the image information merged from the object boundary block merging unit. A shape-adaptive DCT unit performing shape-adaptive DCT by using the merge / separation information input from the image and shape information of the merged image; A quantization unit that receives and quantizes the shape-adaptive DCT coefficient from the quantitative DCT unit, and the merge / separation input from the object boundary block merging / separation information extraction unit by receiving the quantized shape-adaptive DCT coefficient from the quantization unit. A shape-adaptive DCT coefficient and an associated information encoding unit for performing an adaptive scan that applies a zigzag scan by skipping the region where the coefficient does not exist using the information and the shape information of the merged image are encoded.

The video information encoding method of the MPEG-4 encoder according to the second embodiment of the present invention for achieving the above object is to determine whether or not to merge the object boundary block using the restored shape information. The first step of determining and extracting the object boundary block merging / separation information according to the determination, and using the shape information restored from the original image information or the image error information and the merging / separation information extracted in the first step, A second step of merging the image information, a third step of performing shape adaptation DCT by receiving the merged / separated information and the shape information of the merged image from the image information merged in the second step, and the third step; The fourth step of quantizing the shape-adaptive DCT coefficients obtained in the step; and the merged / separated information and the merged zero by receiving the quantized shape-adaptive DCT coefficients obtained in the fourth step. Skip for using the shape information of the area factor does not exist, characterized by comprising a fifth step of encoding by executing the adaptive scan applying the zigzag scanning.

In order to achieve the above object, the image information encoding apparatus of the MPEG-4 encoder according to the third embodiment of the present invention receives whether the object shape block is to be merged and separated from the received shape information. Object boundary block merging / separation information extractor for determining and extracting object boundary block merging / separation information, and shape information restored by receiving original image information or image error information, and object boundary block merging / separation information extracting unit The object boundary block merging unit for merging the image information of the object boundary block by using the merge / separation information inputted from the object and the shape information for padding by receiving merge / separation information from the object boundary block merging / separation information extraction unit. Original image shape information, restored shape information or arbitrary shape information created using original image shape information and restored shape information) Padding shape information extraction unit for outputting the shape information for the padding of the boundary frame block merged and non-merged block, and merged input from the padding shape information extraction unit receiving image information from the object boundary block merger An out-of-object area padding unit for padding the out-of-object area by using shape information for padding the unblocked and unmerged blocks; a DCT unit for receiving DCT from the out-of-object area padding unit and performing DCT; A DCT that receives a DCT coefficient from the DCT unit and quantizes the DCT coefficient, and a DCT variable-coded using the merge / separation information input from the object boundary block merging / separation information extraction unit. And a coefficient and related information encoder.

In order to achieve the above object, the video information encoding method of the MPEG-4 encoder according to the third embodiment of the present invention determines whether to merge or separate object boundary blocks as reconstructed shape information. Image information of the object boundary block using the first step of extracting object boundary block merging / separation information, shape information reconstructed from the original image information or image error information, and merge / separation information extracted in the first step. A second step of merging and the shape information for the merging / separation information and padding extracted in the first step (original image shape information, reconstructed shape information or original image shape information and reconstructed shape information) The shape information for the padding of the merged and non-merged blocks using the shape information, etc.); A fourth step of padding the out-of-object region by using the shape information for padding the merged block and the non-merged block obtained by the merged image information in the third step; and DCT padding the image information padded in the fourth step. A variable length using the merged / separated information extracted in the first step, the sixth step of quantizing the DCT coefficient obtained in the fifth step, and the quantized DCT coefficient obtained in the sixth step. And a seventh step of encoding.

Hereinafter, an image information encoding apparatus and method thereof of an MPEG-4 encoder according to an embodiment of the present invention will be described in detail with reference to the accompanying drawings.

The present invention relates to a novel technique of a video information encoder in the MPEG-4 encoder of FIG.

FIG. 10 is a diagram illustrating an image information encoding apparatus of an MPEG-4 encoder according to a first embodiment of the present invention, and the image information encoding apparatus 1000 of the MPEG-4 encoder according to the first embodiment of the present invention is shown in FIG. As shown in FIG. 10, the out-of-object region padding unit 1100, the object boundary block merging / separation information extraction unit 1200, the object boundary block merging unit 1300, the DCT unit 1400, the quantization unit 1500, and the DCT It is composed of a coefficient and related information encoder 1600, an inverse quantizer 1700, an inverse DCT unit 1800, and a merged block separator 1900.

The out-of-object region padding unit 1100 receives original image information or image error information from the subtractor 500 and shape information (original image shape information, reconstructed) for padding input from the shape information encoder 300. Padding the area outside the object of the image information by using the shape information or the original image shape information and arbitrary shape information generated using the restored shape information.

The object boundary block merging / separation information extracting unit 1200 receives the shape information restored from the shape information encoding unit 300 and determines whether or not to merge the object boundary block and whether to separate the object boundary block. It extracts and outputs object boundary block merge / separation information.

The object boundary block merging unit 1300 receives the image information of the out-of-object area padding unit 1100 and extracts the restored shape information input from the shape information encoding unit 300 and the object boundary block merging / separation information. The merger / separation information input from the unit 1200 serves to merge the image information of the object boundary block.

The DCT unit 1400 receives the image information from the object boundary block merging unit 1300 to perform DCT, and the quantization unit 1500 receives the DCT coefficient from the DCT unit 1400 to quantize it. The DCT coefficient and the related information encoder 1600 may receive the DCT coefficient quantized by the quantizer 1500 and merge / separate the object boundary block merge / separation information extractor 1200. After variable-length encoding by using the information and outputs to the multiplexer 900.

The inverse quantization unit 1700 receives the quantized signal from the quantization unit 1500 and performs inverse quantization, and the inverse DCT unit 1800 inputs the inverse quantization signal from the inverse quantization unit 1700. Receiving and performing a reverse DCT, and the merged block separator 1900 receives a signal decoded by the reverse DCT unit 1800 and is input from the object boundary block merger / separation information extractor 1200. It divides the merged blocks using merge / separation information and outputs reconstructed image information to the VOP detector 800 through the adder 700.

An encoding method using the image information encoding apparatus of the MPEG-4 encoder according to the first embodiment of the present invention configured as described above will be described.

First, when the object outside area padding unit 1100 receives the original image information or the image error information from the subtractor 500, the shape information for the padding from the shape information encoder 300 (the original image shape information and the restored image) is received. The shape information or the original image shape information and arbitrary shape information generated using the reconstructed shape information, etc.) are input to pad the area outside the object.

Subsequently, the object boundary block merging unit 1300 receives image information padded by the out-of-object region padding unit 1100 and merges / separates information and shape information extracted by the object boundary block merging / separation information extracting unit 1200. The image information of the object boundary block is merged using the reconstructed shape information input from the encoder 300.

In this case, the DCT unit 1400 receives the image information from the object boundary block merging unit 1300 to perform DCT, and the quantization unit 1500 receives the DCT coefficient from the DCT unit 1400 and quantizes it.

Subsequently, the DCT coefficient and the related information encoder 1600 receive the DCT coefficient quantized by the quantizer 1500 and receive merge / separation information input from the object boundary block merge / separation information extractor 1200. The variable length encoding is performed and output to the multiplexer 900.

Meanwhile, the inverse quantization unit 1700 receives inverse quantization by receiving the quantized signal from the quantization unit 1500, and the inverse DCT unit 1800 receives the inverse quantized signal from the inverse quantization unit 1700. The merged block separator 1900 receives a signal decoded by the inverse DCT unit 1800 and uses merge / separation information input from the object boundary block merge / separation information extractor 1200. After the merged blocks are separated, the reconstructed image information is output to the VOP detector 800 through the adder 700.

An operation of extracting merge / separation information from the object boundary block merge / separation information extractor 1200 will be described in detail.

12 is a diagram showing a block arrangement in a macroblock according to the video format of a video to be encoded, (a) showing four 8x8 blocks in a macroblock, and (b) in a macroblock. The luminance block (Y) and chrominance block (Cb, Cr) ratio arranged in the ratio is 4: 2: 0, (c) is the luminance block (Y) and chrominance block (Cb, Cr) ratio arranged in the macroblock is 4 2: 2, and (d) shows that the ratio of the luminance block Y and the color difference blocks Cb and Cr arranged in the macroblock is 4: 4: 4.

FIG. 13 is a diagram illustrating a model in which object boundary block merging / separation is applied to each image format of FIG. 12, where (a) shows four 8 × 8 blocks in a macroblock, and (b) shows a macroblock. (C) shows the case where the ratio is 4: 2: 2, and (d) shows the case where the ratio is 4: 4: 4.

First, the horizontal merger information extraction process in the horizontal merger, vertical merger, diagonal merger information extraction process of the object boundary block will be described.

In the first step, if block 1 and block 2 are object boundary blocks, and the shape information of block 2 is rotated 180 degrees and there are no pixels overlapping each other among the pixels in the object between the shape information of block 1, the image information of block 2 ( Original image information or error information between images) is rotated 180 degrees to merge with the image information of block 1, and block 2 extracts a flag to be treated as an out-of-object block by the image information encoding apparatus.

In the second step, if block 3 and block 4 are object boundary blocks, and the shape information of block 4 is rotated 180 degrees and there are no pixels overlapping each other among the pixels in the object between the shape information of block 3, the image information of block 4 ( Original image information or inter-image error information) by 180 degrees to be merged into the image information of block 3, and block 4 extracts a flag to be treated as an out-of-object block by the image information encoding apparatus.

In addition, the vertical merger information extraction process is described as follows.

In the first step, if block 1 and block 3 are object boundary blocks, and the shape information of block 3 is rotated 180 degrees and there are no pixels overlapping each other among the pixels in the object between the shape information of block 1, the image information of block 3 ( Original image information or error information between images) is rotated 180 degrees to be merged into the image information of block 1, and block 3 extracts a flag to be treated as an out-of-object block by the image information encoding apparatus.

In the second step, if block 2 and block 4 are object boundary blocks, and the shape information of block 4 is rotated 180 degrees and there are no pixels overlapping each other among the pixels in the object between the shape information of block 2, the image information of block 4 ( Original image information or error information between images) is rotated 180 degrees to be merged into the image information of block 2, and block 4 extracts a flag to be treated like an out-of-object block by the image information encoding apparatus.

In addition, the diagonal merge information extraction process is described as follows.

In the first step, if block 1 and block 4 are object boundary blocks, and the shape information of block 4 is rotated 180 degrees and there are no pixels overlapping each other among the pixels in the object between the shape information of block 1, the image information of block 4 ( Original image information or error information between images) is rotated 180 degrees to merge with the image information of block 1, and block 4 extracts a flag to be treated as an out-of-object block by the image information encoding apparatus.

In the second step, if block 2 and block 3 are object boundary blocks, and the shape information of block 3 is rotated 180 degrees and there are no pixels overlapping each other among the pixels in the object between the shape information of block 2, the image information of block 3 ( Original image information or error information between images) is rotated 180 degrees to merge with the image information of block 2, and block 3 extracts a flag to be treated like an out-of-object block by the image information encoding apparatus.

The object boundary block merging information extraction process according to the present invention may be applied in one of the following orders in the horizontal, vertical and diagonal merging information extraction process as described above.

1. Horizontal merger information extraction → vertical merger information extraction → diagonal merger information extraction

2. Horizontal merger information extraction → diagonal merger information extraction → vertical merger information extraction

3. Vertical Merger Information Extraction → Horizontal Merger Information Extraction → Diagonal Merger Information Extraction

4. Extract vertical merger information → Extract diagonal merger information → Extract horizontal merger information

5. Diagonal merger information extraction → Vertical merger information extraction → Horizontal merger information extraction

6. Diagonal merger information extraction → Horizontal merger information extraction → Vertical merger information extraction

On the other hand, the merged information extraction process is applied differently as shown in Figure 13 according to the image format (4: 2: 0, 4: 2: 2 :, 4: 4: 4) as shown in FIG.

(a). When there is only a luminance block without a chrominance block or consists of four blocks such as transparency information of grayscale (GRAYSCALE) shape information, the merge information extraction processes 1 to 6 may be applied in any order.

(b). In the case of 4: 2: 0 format, the luminance blocks may be applied in one of the merge information extraction processes 1 to 6 above.

(c). In the case of 4: 2: 2 format, the luminance blocks may be applied in any one of the merged information extraction processes 1 to 6, and the color difference blocks may be applied only to the horizontal merged information extraction process by color difference components (Cb and Cr). have.

(d). In the case of 4: 4: 4 format, the luminance blocks may be applied in one order of the merge information extraction processes 1 to 6, and the chrominance blocks may be applied to each of the merge information extraction processes 1 to 6 by color difference components (Cb and Cr). It can be applied in either order.

Meanwhile, the merging information extraction process in the object boundary block merging / separation information extracting units 1200 and 2200 of FIGS. 10 and 11 is merged in the process of moving the arrow (→) in the merging information extraction process (1 to 6). Is extracted, the process after the arrow (→) is omitted, and to the object boundary block merging unit 1300 and the merged block separation unit 1900 so that merging and separation occurs in the course of the flag. Output

On the other hand, the object boundary block merging unit 1300 receives a flag related to the merged information extracted from the object boundary block merging / separation information extracting unit 1200 and shape information restored from the shape information encoding unit 300. An object boundary block merging operation of image information, which is a macroblock unit received by the object outside area padding unit 1100, will be described.

FIG. 14 is a flowchart showing the operation of the object boundary block merging unit according to the first embodiment of the present invention, where (a) is an overall operation flowchart performed in the object boundary block merging unit, and (b) is a diagram of FIG. This is an operation flowchart applied to all of the steps S1 and S3 in (a).

First, the object boundary block merging unit 1300 applies the object boundary block merger to the luminance block of the image information which is a macroblock unit received from the object padding region 1100 (S1).

Subsequently, it is determined whether the color difference block exists in the input image information and the image format is not 4: 2: 0 (S2).

If the format of the input image information is 4: 2: 2 or 4: 4: 4 in step S2 (YES), the object boundary block merger is applied to the color difference block for each color difference component (S3).

On the other hand, if the chrominance block does not exist or the image format is 4: 2: 0 in step S2 (NO), the process ends.

In the step S1, the application of the object boundary block merger to the luminance block of the image information in macroblock units will be described with reference to FIG. 14B.

First, the object boundary block merging unit 1300 merges / separates information of one block Bi among four 8 × 8 blocks in an input 16 × 16 luminance block, and then merges / separates the information boundary unit 1200 of the object boundary block. It is determined whether or not the block (Bi) is a block to merge the other block is received from the (S1-1).

If, in step S1-1, if any one of the blocks Bi is a merged block (YES), as shown in Fig. 16, the block Bi and a block merged with the block Bi are shown. Merger (S1-2).

FIG. 16 is a diagram illustrating an object boundary block merging technology according to a first embodiment of the present invention, in which (a) shows a block before merging, in which a block to be merged is an upper left block, and a block to be merged is an upper right corner. It is assumed that the block is a block, the outside object pixel value of the upper left block is a, and the padded outside object pixel value of the upper right block is b.

(b) shows a block after the merge, and the detailed process of merging first rotates the merged block by 180 degrees and then merges the merged block. In this case, the image information of the intra-object region of the merged block rotated 180 degrees replaces the padded information of the merged block, and the padded information of the extra-object region of the merged block rotated 180 degrees is padded information of the merged block. And the averaged value {(a + b) / 2}.

Subsequently, after increasing the number of i by one (S1-3), it is determined whether i is less than 4 (S1-4), and if i is less than 4 (YES), the step (S1-) While proceeding to 1), when i reaches 4, it ends.

On the other hand, in step S1-1, if the block Bi is not a block for merging other blocks, that is, a block to be merged or an out-of-object block NO, the flow advances to step S1-3.

Meanwhile, a process of merging an object boundary block with an arbitrary color difference block by color difference components Cb and Cr in step S3 will be described with reference to FIG. 14B.

First, the object boundary block merging unit 1300 merges / separates merge / separation information of one block Bi among two 8 × 8 blocks in color difference information Cb or Cr of the input image information. In operation S3-1, the information extracting unit 1200 receives the input from the information extracting unit 1200 and determines whether the block Bi is a block for merging other blocks.

In step S3-1, if any one of the blocks Bi is a merged block (YES), the block Bi and a block merged with the block are merged as shown in Fig. 16 (S3-). 2).

FIG. 16 is a diagram illustrating an object boundary block merging technology according to a first embodiment of the present invention, in which (a) shows a block before merging, in which a block to be merged is an upper left block, and a block to be merged is an upper right corner. It is assumed that the block is a block, the outside object pixel value of the upper left block is a, and the padded outside object pixel value of the upper right block is b.

(b) shows a block after the merge, and the detailed process of merging first rotates the merged block by 180 degrees and then merges the merged block. In this case, the image information of the intra-object region of the merged block rotated 180 degrees replaces the padded information of the merged block, and the padded information of the extra-object region of the merged block rotated 180 degrees is padded information of the merged block. And the averaged value {(a + b) / 2}.

Subsequently, after increasing the number of i by one (S3-3), it is determined whether i is less than 2 (S3-4), and when i is less than 2 (YES), the step (S3-) On the other hand, the process proceeds to 1) and ends when i becomes 2.

On the other hand, in step S3-1, if the block Bi is not a block for merging other blocks, that is, a block to be merged or an out-of-object block NO, the flow advances to step S3-3.

Meanwhile, the merged block separator 1900 receives a signal decoded by the inverse DCT unit 1800 and merges the merged data using the merge / separation information input from the object boundary block merge / separation information extractor 1200. A technique of reconstructing image information by separating the blocks is described in detail.

FIG. 15 is a flowchart showing the operation of the object boundary block separator according to the first and second embodiments of the present invention, and (a) is a diagram showing an overall operation flowchart performed in the merged block separator. ) Is an operational flowchart applied to both steps S1 and S3 in FIG.

First, the merged block separator 1900 applies the object boundary block separation to the luminance block of the image information that is the macroblock unit received from the inverse DCT unit 1800 (S1).

Subsequently, it is determined whether the color difference block exists in the input image information and the image format is not 4: 2: 0 (S2).

If the format of the input image information is 4: 2: 2 or 4: 4: 4 at step S2 (YES), object boundary block separation is applied to the color difference block by color difference components (Cb, Cr). (S3).

On the other hand, if the chrominance block does not exist or the image format is 4: 2: 0 in step S2 (NO), the process ends.

In the step S1, the application of the object boundary block separation to the luminance block of the image information in macroblock units will be described with reference to FIG. 15B.

First, the merged block divider 1900 merges / separates information of one block Bi among four 8 × 8 blocks in an input 16 × 16 luminance block by the object boundary block merged / separated information extractor 1200. ), It is determined whether the block Bi merges with another block (S1-1).

If, in step S1-1, if any one of the blocks Bi is a merged block (YES), the merged block is separated from the block Bi as shown in Fig. 17 ( S1-2).

17 is a view showing the object boundary block separation technology according to the first and second embodiments of the present invention, (a) shows a merged block before separation, the merged block is the upper left block, (b ) Shows a block after separation, and the detailed separation process first identifies a merged block based on merge / separation information when it is determined to be a merged block, and 180 degrees of the merged block by image information of the identified block. Replace all with rotated image information. Although the separated state in FIG. 17B seems strange, unnecessary information is ignored by the shape information in the merging process, so that only the copying process is performed.

Subsequently, after increasing the number of i by one (S1-3), it is determined whether i is less than 4 (S1-4), and if i is less than 4 (YES), the step (S1-) On the other hand, the process proceeds to 1) and terminates when i becomes 4.

On the other hand, in step S1-1, if the block Bi is not a block in which another block is merged (NO), the process proceeds to the step S1-3.

Meanwhile, a process of separating the object boundary block from an arbitrary color difference block by color difference component Cb and Cr in step S3 will be described with reference to FIG. 15B.

First, the merged block separator 1900 merges / separates information of one block Bi among two 8 × 8 color difference blocks in two color difference blocks of the input image information. It is determined in step 1200 whether the block Bi merges with another block (S3-1).

In step S3-1, if any one of the blocks Bi is a merged block (YES), the merged block is separated from the block Bi as shown in Fig. 17 (S3-). 2).

17 is a view showing the object boundary block separation technology according to the first and second embodiments of the present invention, (a) shows a merged block before separation, the merged block is the upper left block, (b ) Shows a block after separation, and the detailed separation process first identifies a merged block based on merge / separation information when it is determined to be a merged block, and 180 degrees of the merged block by image information of the identified block. Replace all with rotated image information. Although the separated state in FIG. 17B seems strange, unnecessary information is ignored by the shape information in the merging process, so that only the copying process is performed.

Subsequently, after increasing the number of i by one (S3-3), it is determined whether i is less than 2 (S3-4), and when i is less than 2 (YES), the step (S3-) On the other hand, the process proceeds to 1) and ends when i becomes 2.

On the other hand, in step S3-1, if the block Bi is not a block in which another block is merged (NO), the process proceeds to the step S3-3.

FIG. 11 is a diagram illustrating an image information encoding apparatus of an MPEG-4 encoder according to a second embodiment of the present invention, and the image information encoding apparatus 2000 of the MPEG-4 encoder according to the second embodiment of the present invention is shown in FIG. As shown in Fig. 11, the object boundary block merging / separation information extraction unit 2200, the object boundary block merging unit 2100, the shape adaptation DCT unit 2300, the quantization unit 2400, the shape adaptation DCT coefficient, and related information The information encoding unit 2500, the inverse quantization unit 2600, the inverse shape adaptive DCT unit 2700, and the merged block separation unit 2800 are constituted.

The object boundary block merging / separation information extracting unit 2200 receives the shape information restored from the shape information encoding unit 300 and determines whether or not to merge the object boundary block and whether to separate the object boundary block. It extracts and outputs object boundary block merge / separation information.

The object boundary block merging unit 2100 receives the original image information or the image error information from the subtractor 500 and the restored shape information and the object boundary block merging / separation information received from the shape information encoding unit 300. The merger / separation information input from the extraction unit 2200 serves to merge the image information of the object boundary block.

The shape-adaptive DCT unit 2300 receives image information from the object boundary block merging unit 2100 and merges / separates information input from the object boundary block merging / separation information extracting unit 2200 and the merged image. It receives shape information and performs shape adaptation DCT.

The quantization unit 2400 serves to quantize the shape-adaptive DCT coefficient from the shape-adaptive DCT unit 2300.

The shape-adaptive DCT coefficient and the related information encoder 2500 receive the shape-adaptive DCT coefficients quantized by the quantization unit 2400 and merge the inputs from the object boundary block merging / separation information extraction unit 2200. Using the separation information and the shape information of the merged image, it skips over an area where coefficients do not exist and executes an adaptive scan that applies a zigzag scan and encodes it to output to the multiplexer 900. do.

The inverse quantizer 2600 serves to inverse quantize the signal quantized by the quantizer 2400.

The inversely adaptive DCT unit 2700 receives the inverse quantized signal from the inverse quantization unit 2600 and merges / separates information input from the object boundary block merging / separation information extracting unit 2200 and merged shape. It plays a role of performing inverse-adaptive DCT using information.

The merged block separator 2800 receives a signal decoded by the inversely adaptive DCT unit 2700 and uses merge / separation information input from the object boundary block merge / separation information extractor 2200. And separates the merged block and outputs the reconstructed image information to the VOP detector 800 through the adder 700.

The encoding method using the image information encoding apparatus of the MPEG-4 encoder according to the second embodiment of the present invention configured as described above will be described.

First, the object boundary block merging / separation information extracting unit 2200 receives shape information restored from the shape information encoding unit 300 and determines whether or not to merge the object boundary block and whether to separate the object boundary block. Extract object boundary block merge / separation information.

Subsequently, the object boundary block merging unit 2100 receives original image information or image error information from the subtractor 500 and merges / separates the restored shape information and the object boundary block inputted from the shape information encoder 300. The image information of the object boundary block is merged using the merge / separation information input from the information extractor 2200.

Also, the shape-adaptive DCT unit 2300 receives image information from the object boundary block merging unit 2100 and merges / separates information input from the object boundary block merging / separation information extracting unit 2200 and the merged image. Perform shape adaptation DCT by receiving shape information of.

In addition, the quantization unit 2400 receives the quantization of the shape-adaptive DCT coefficient from the shape-adaptive DCT unit 2300, and the shape-adaptive DCT coefficient and the related information encoder 2500 are quantized by the quantization unit 2400. By receiving the quantized shape-adaptive DCT coefficient and skipping the region where the coefficient does not exist using the merge / separation information input from the object boundary block merge / separation information extractor 2200 and the shape information of the merged image By performing an adaptive scan (ADAPTIVE SCAN) to apply a zigzag scan to be encoded and output to the multiplexer 900.

Meanwhile, the inverse quantization unit 2600 receives inverse quantization by receiving the quantized signal from the quantization unit 2400, and the inverse-adaptive DCT unit 2700 inputs the inverse quantization signal from the inverse quantization unit 2600. In response to the object boundary block merger / separation information extraction unit 2200, the merge / separation information and the merged shape information are input to perform an inverse shape adaptive DCT.

Subsequently, the merged block separator 2800 receives the signal decoded by the inversely-adaptive DCT unit 2700, and receives merge / separation information input from the object boundary block merge / separation information extractor 2200. The merged blocks are separated, and the reconstructed image information is output to the VOP detector 800 through the adder 700.

The process of extracting the object boundary block merging / separation information using the shape information restored by the object boundary block merging / separation information extracting unit 2200 is the same as in the image information encoding method to which padding technology is applied.

In addition, the object boundary block merging process performed by the object boundary block merging unit 2100 is almost similar to the image information incubation method to which the padding technology is applied. The image information (original image information or image error information) is rotated 180 degrees and merged into the merged block. At this time, the image information of the in-object region of the merged block rotated 180 degrees indicates information of the out-of-object region of the merged block. All information in the out-of-object area to be merged and merged and the out-of-object area to be merged is the same as in the case of padding technology.

In addition, the merged block separation operation performed by the merged block separator 2800 is the same as the image information encoding method to which the padding technique is applied.

FIG. 18 is a diagram showing an image information encoding apparatus of an MPEG-4 encoder according to a third embodiment of the present invention, and the image information encoding apparatus of the MPEG-4 encoder according to the third embodiment of the present invention is shown in FIG. As described above, the object boundary block merging / separation information extracting unit 3200, the object boundary block merging unit 3100, the padding shape information extracting unit 3300, the object boundary region padding unit 3400, the DCT unit 3500, and quantization A unit 3600, a DCT coefficient and related information encoder 3700, an inverse quantizer 3800, an inverse DCT unit 3900, and a merged block separator 4000 are provided.

The object boundary block merging / separation information extracting unit 3200 receives shape information restored from the shape information encoding unit 300 and determines whether or not to merge the object boundary block and whether to separate the object boundary block. It extracts and outputs object boundary block merge / separation information.

The object boundary block merging unit 3100 receives the original image information or the image error information from the subtractor 500 and the restored shape information and the object boundary block merging / separation information input from the shape information encoding unit 300. The merger / separation information input from the extractor 3200 serves to merge image information of the object boundary block.

The padding shape information extractor 3300 receives merger / separation information from the object boundary block merger / separation information extractor 3200 and shape information for padding from the shape information encoder 300 (original image shape). Information, restored shape information or arbitrary shape information created using original shape information and restored shape information, etc.), and outputs shape information for padding of the object boundary block and the unmerged block. It plays a role.

The object outside area padding unit 3400 receives image information from the object boundary block merging unit 3100 to form padding of merged blocks and non-merged blocks input from the padding shape information extracting unit 3300. Padding the area outside the object by using the information.

The DCT unit 3500 receives image information from the object outside area padding unit 3400 to perform DCT, and the quantization unit 3600 receives the DCT coefficient from the DCT unit 3500 to quantize the quantization unit. The DCT coefficient and the related information encoder 3700 may receive a DCT coefficient quantized by the quantization unit 3600 and merge / separate the object boundary block merge / separation information extractor 3200. Variable length encoding by using the information to output to the multiplexer 900.

The inverse quantization unit 3800 receives inverse quantization by receiving the quantized signal from the quantization unit 3600, and the inverse DCT unit 3900 inputs the inverse quantized signal in the inverse quantization unit 3800. And performs a reverse DCT, and the merged block separator 4000 receives a signal decoded by the reverse DCT unit 3900 and is input from the object boundary block merge / separation information extractor 3200. It divides the merged blocks using merge / separation information and outputs reconstructed image information to the VOP detector 800 through the adder 700.

An encoding method using the image information encoding apparatus of the MPEG-4 encoder according to the third embodiment of the present invention configured as described above will be described.

First, the object boundary block merging / separation information extracting unit 3200 receives shape information restored from the shape information encoding unit 300 and determines whether or not to merge the object boundary block and whether to separate the object boundary block. Extracts the object boundary block merge / separation information accordingly and outputs

Subsequently, the object boundary block merging unit 3100 receives original image information or image error information from the subtractor 500 and merges / separates the restored shape information and the object boundary block inputted from the shape information encoder 300. The image information of the object boundary block is merged using the merge / separation information input from the information extracting unit 3200.

In this case, the padding shape information extractor 3300 receives merger / separation information from the object boundary block merger / separation information extractor 3200 and shape information for padding from the shape information encoder 300 (the original image). Shape information, restored shape information or arbitrary shape information created using original shape information and restored shape information), etc.) and inputs shape information for padding of the object boundary block and the unmerged block. Output

Subsequently, the object outside area padding unit 3400 receives image information from the object boundary block merging unit 3100 to pad the merged block and the unmerged block that are input from the padding shape information extracting unit 3300. Pad the area outside the object using shape information.

In addition, the DCT unit 3500 receives the image information from the outside object padding unit 3400 and performs DCT, and the quantization unit 3600 receives the DCT coefficient from the DCT unit 3500 and quantizes it.

Subsequently, the DCT coefficient and related information encoder 3700 receives the DCT coefficient quantized by the quantizer 3600 and uses the merge / separation information input from the object boundary block merge / separation information extractor 3200. Variable length encoding is output to the multiplexer 900.

On the other hand, the inverse quantizer 3800 receives the quantized signal from the quantizer 3600 and inverse quantizes, and the inverse DCT unit 3900 receives the inverse quantized signal from the inverse quantizer 3800 and receives inverse DCT. Perform

Subsequently, the merged block separating unit 4000 receives the signal decoded by the inverse DCT unit 3900 and merges using the merge / separation information input from the object boundary block merging / separation information extracting unit 3200. The block is divided and outputs the reconstructed image information to the VOP detector 800 through the adder 700.

The merging / separating information extraction process in the object boundary block merging / separation information extracting unit 3200 is the same as in the first and second embodiments.

The object boundary block merging unit 3100 receives a flag relating to the merged information extracted from the object boundary block merging / separation information extracting unit 3200 and shape information restored from the shape information encoding unit 300 and uses the received information. The object boundary block merging operation of the original image information or the image error information input in macroblock units by the subtractor 500 will be described.

FIG. 19 is a flowchart showing the operation of the object boundary block merging unit according to the third embodiment of the present invention, where (a) is an overall operation flowchart performed in the object boundary block merging unit, and (b) is a diagram of FIG. This is an operation flowchart applied to all of the steps S1 and S3 in (a).

First, the object boundary block merging unit 3100 applies the object boundary block merger to the luminance block of the image information which is the macro block unit input from the subtractor 500 (S1).

It is determined whether a color difference block exists in the input image information and whether the image format is not 4: 2: 0 (S2).

If the format of the input image information is 4: 2: 2 or 4: 4: 4 in step S2 (YES), the object boundary block merger is applied to the color difference block for each color difference component (S3).

On the other hand, if the chrominance block does not exist or the image format is 4: 2: 0 in step S2 (NO), the process ends.

The application of the object boundary block merger to the luminance block of the image information in macroblock units in step S1 will be described with reference to FIG. 19B.

First, the object boundary block merging unit 3100 merges / separates information of one block Bi among four 8 × 8 blocks in an input 16 × 16 luminance block, and then merges / separates information extracting unit 3200 the object boundary block. It is determined whether or not the block (Bi) is a block to merge the other block is received from the (S1-1).

If the arbitrary one block Bi is a merged block (YES) in step S1-1, the block Bi and a block merged with the block are merged as shown in Fig. 21 (S1-). 2).

21 is a view showing the object boundary block merging technology according to the third embodiment of the present invention, where (a) shows a block before merging, and the block to be merged is the upper left block and the block to be merged is the upper right Assume a block (b) shows a block after the merge, and the detailed process of merging first rotates the merged block by 180 degrees and then merges the merged block. In this case, the image information of the intra-object region of the merged block rotated 180 degrees replaces the information of the merged block.

Subsequently, after increasing the number of i by one (S1-3), it is determined whether i is less than 4 (S1-4), and if i is less than 4 (YES), the step (S1-) While proceeding to 1), when i reaches 4, it ends.

On the other hand, in step S1-1, if the block Bi is not a block for merging other blocks, that is, a block to be merged or an out-of-object block (NO), the process proceeds to step S1-3.

Meanwhile, a process of merging an object boundary block with an arbitrary color difference block by color difference components Cb and Cr in step S3 will be described with reference to FIG. 19B.

First, the object boundary block merging unit 3100 merges / separates the object boundary block merging / separation information for one block Bi among two 8 × 8 blocks in the color difference information Cb or Cr of the input image information. In operation S3-1, the information extracting unit 3200 receives the input from the information extracting unit 3200 and determines whether the block Bi is a block for merging other blocks.

If the arbitrary block Bi is a merged block in step S3-1 (YES), the block Bi and a block merged with the block are merged as shown in Fig. 21 (S3-). 2).

21 is a view showing the object boundary block merging technology according to the third embodiment of the present invention, in which (a) shows a block before merging, and the block to be merged is the upper left block, and the block to be merged is the upper right. It is a block. (b) shows a block after the merge, and the detailed process of merging first rotates the merged block by 180 degrees and then merges the merged block. In this case, the image information of the intra-object region of the merged block rotated 180 degrees replaces the information of the merged block.

Subsequently, after increasing the number of i by one (S3-3), it is determined whether i is less than 2 (S3-4), and when i is less than 2 (YES), the step (S3-) On the other hand, the process proceeds to 1) and ends when i becomes 2.

On the other hand, in step S3-1, if the block Bi is not a block for merging other blocks, that is, a block to be merged or an out-of-object block NO, the flow advances to step S3-3.

Meanwhile, the merged block separator 4000 receives a signal decoded by the inverse DCT unit 3900 and merges the merged data using the merge / separation information input from the object boundary block merge / separation information extractor 3200. A technique of reconstructing image information by separating the blocks is described in detail.

FIG. 20 is a flowchart showing the operation of the object boundary block separator according to the third embodiment of the present invention, where (a) is an overall operation flowchart performed in the merged block separator, and (b) This is an operational flowchart applied to both steps S1 and S3 in FIG.

First, the merged block separator 4000 applies the object boundary block separation to the luminance block of the image information which is the macroblock unit input from the inverse DCT unit 3900 (S1).

It is determined whether a color difference block exists in the input image information and whether the image format is not 4: 2: 0 (S2).

In step S2, when the format of the input image information is 4: 2: 2 or 4: 4: 4 (YES), object boundary block separation is applied to the color difference block for each color difference component.

On the other hand, if the chrominance block does not exist or the image format is 4: 2: 0 in step S2 (NO), the process ends.

In the step S1, the application of the object boundary block separation to the luminance block of the image information in macroblock units will be described with reference to FIG. 20B.

First, the merged block separating unit 4000 merges / separates information of one block Bi among four 8 × 8 blocks in an input 16 × 16 luminance block by the object boundary block merging / separation information extracting unit 3200. ), It is determined whether the block Bi merges with another block (S1-1).

If the one arbitrary block Bi is a block in which the other blocks are merged in the step S1-1 (YES), the merged blocks are separated from the block Bi as shown in Fig. 22 (S1-). 2).

22 is a view showing the object boundary block separation technology according to the third embodiment of the present invention, (a) shows a merged block before separation, and the merged block is the upper left block, and (b) The detailed process for separating, which shows the block after separation, first determines the merged block by the merge / separation information when it is determined that the merged block, and rotates the merged block 180 degrees by the image information of the identified block. Replace all with video information. Although the separated state in FIG. 22 (b) seems strange, unnecessary information is ignored by the shape information in the merging process, so that only the copying process is performed.

Subsequently, after increasing the number of i by one (S1-3), it is determined whether i is less than 4 (S1-4), and if i is less than 4 (YES), the step (S1-) While proceeding to 1), when i reaches 4, it ends.

On the other hand, in step S1-1, if the block Bi is not a block in which another block is merged (NO), the process proceeds to the step S1-3.

Meanwhile, a process of separating the object boundary block from any color difference block by the color difference components Cb and Cr in step S3 will be described with reference to FIG. 20B.

First, the merged block separating unit 4000 merges / separates information of one block Bi among two 8 × 8 color difference blocks in two color difference blocks of the input image information, and merges / separates information extracting unit of the object boundary block. In operation S3-1, it is determined whether the block Bi is merged with another block by receiving the input at 3200.

In step S3-1, if any one of the blocks Bi is a merged block (YES), the merged block is separated from the block Bi as shown in Fig. 22 (S3-). 2).

22 is a view showing the object boundary block separation technology according to the third embodiment of the present invention, (a) shows a merged block before separation, and the merged block is the upper left block, and (b) The detailed process for separating, which shows the block after separation, first determines the merged block by the merge / separation information when it is determined that the merged block, and rotates the merged block 180 degrees by the image information of the identified block. Replace all with video information. Although the separated state in FIG. 22 (b) seems strange, unnecessary information is ignored by the shape information in the merging process, so that only the copying process is performed.

Subsequently, after increasing the number of i by one (S3-3), it is determined whether i is less than 2 (S3-4), and when i is less than 2 (YES), the step (S3-) On the other hand, the process proceeds to 1) and ends when i becomes 2.

On the other hand, in step S3-1, if the block Bi is not a block in which another block is merged (NO), the process proceeds to the step S3-3.

On the other hand, in the video information encoding method of the MPEG-4 encoder according to the first, second, and third embodiments, i-values for comparing the object boundary block merging unit and the merged block separating unit during merging and separating color difference blocks are compared. Although it is set to 2, this value is applied when the image format is assumed to be 4: 2: 2. If the image format is 4: 4: 4, the i value is 4.

As described above, according to the video information encoding apparatus and method of the MPEG-4 encoder according to the present invention, it is determined by using the reconstructed shape information whether or not to merge or separate object boundary blocks when encoding region-based image information. After that, the object boundary block merging / separation information extraction unit extracting the object boundary block merging / separation information is separated from the object boundary block merging unit and the merged block separation unit, thereby avoiding the redundant configuration. It is an excellent effect that can be reduced.

1 is a hardware block diagram of a typical MPEG-4 video encoder;

FIG. 2 is a view showing a VOP, which is an image composed only of information in a whole image or a meaningful object region constituting the image in an arbitrary image;

3 is a diagram illustrating the type of macroblocks in a VOP;

4 shows block types in a VOP;

FIG. 5 is a view for explaining a low frequency component extended padding technique in a region-based image encoding method; FIG.

6 is a diagram for describing a zero padding technique in an area-based image encoding method;

7 is a view for explaining the shape-adaptive DCT technology in the area-based image encoding method;

8 is a block diagram of an image information encoder according to an object boundary block merge / separation technique to which a conventional padding technique is applied;

9 is a block diagram of an image information encoding unit according to an object boundary block merge / separation technique to which a conventional shape-adaptive DCT is applied;

10 is a diagram showing an image information encoding apparatus of an MPEG-4 encoder according to the first embodiment of the present invention;

11 is a diagram showing an image information encoding apparatus of an MPEG-4 encoder according to a second embodiment of the present invention;

12 is a diagram showing a block arrangement in a macroblock according to an image format of an image to be encoded;

FIG. 13 is a diagram illustrating a model to which object boundary block merging / separation is applied to each image format of FIG. 12. FIG.

14 is a flowchart of the operation of the object boundary block merging unit according to the first embodiment of the present invention;

15 is a flowchart showing the operation of the object boundary block separator according to the first and second embodiments of the present invention;

FIG. 16 is a view illustrating an object boundary block merging technology according to a first embodiment of the present invention; FIG.

17 is a view showing a boundary object separation technology according to the first and second embodiments of the present invention;

18 is a diagram showing an image information encoding apparatus of an MPEG-4 encoder according to a third embodiment of the present invention;

19 is a flowchart of the operation of the object boundary block merging unit according to the third embodiment of the present invention;

20 is a flowchart illustrating an operation of an object boundary block separator according to a third embodiment of the present invention;

21 is a view showing the object boundary block merging technology according to a third embodiment of the present invention;

22 is a view illustrating a boundary object separation technique according to a third embodiment of the present invention.

<Description of Symbols for Main Parts of Drawings>

1000: Image information encoding apparatus 1100: Area padding outside the object

1200: object boundary block merger / separation information extraction unit 1300: object boundary block merger

1400: DCT unit 1500: quantization unit

1600: DCT coefficient and related information encoder 1700: Inverse quantizer

1800: reverse DCT unit 1900: merged block separation unit

Claims (49)

An image information encoding apparatus for receiving original image information or image error information and encoding the same using shape information. An off-object area padding unit which receives original image information or image error information and pads an off-object area of the image information by using shape information for padding; An object boundary block merging / separation information extraction unit for determining whether or not to merge the object boundary blocks by receiving the restored shape information and extracting the object boundary block merging / separation information accordingly; The object boundary that merges the image information of the object boundary block by using the shape information restored by receiving the image information padded by the out-of-object region padding unit and the merge / separation information input from the object boundary block merging / separation information extraction unit. Block merger, A DCT unit which receives the image information from the object boundary block merging unit and performs DCT; A quantization unit configured to receive and quantize the DCT coefficients from the DCT unit; The MPEG-4 encoder comprising a DCT coefficient and a related information encoder for variable length encoding by receiving the quantized signal from the quantization unit and using the merge / separation information input from the object boundary block merge / separation information extractor. Image information encoding apparatus. The method of claim 1, And the shape information for padding is original image shape information, reconstructed shape information, or arbitrary shape information generated using the original image shape information and the reconstructed shape information. The method of claim 1, An inverse quantizer configured to receive and inverse quantize the quantized signal from the quantizer; An inverse DCT unit configured to receive an inverse quantized signal from the inverse quantizer and perform inverse DCT to output a decoded signal; The merged block separator which receives the signal decoded by the inverse DCT unit and separates the merged block using the merge / separation information input from the object boundary block merge / separation information extractor to output reconstructed image information And an image information encoding apparatus of an MPEG-4 encoder. A VOP forming unit for forming a VOP, meaning a content object having an arbitrary shape that can be accessed and edited by a user at a point in time on the time axis; A shape information encoder which receives the shape information of the VOP from the VOP forming unit, encodes, decodes and outputs the original image shape information which is currently input; The image information of the VOP formed by the VOP forming unit is received, the image information of the previous VOP reproduced is received, and the shape information encoder receives the shape information or original image shape information reconstructed by the shape information encoder to estimate the motion in macroblock units. A motion estimation unit for outputting motion information, A motion compensator for receiving motion information from the motion estimation unit, receiving shape information reconstructed by the shape information encoder, and outputting image information of the motion-compensated VOP by receiving the reproduced image information of the previous VOP; , A subtractor for receiving the image information of the VOP whose motion is compensated by the motion compensator and obtaining the difference between the two image information by receiving the image information of the VOP formed by the VOP forming unit;  An out-of-object region padding unit which receives the original image information or the image error information from the subtractor and pads the out-of-object region of the image information by using shape information for padding; The object boundary block merging / separation information which receives the shape information restored by the shape information encoder and determines whether to merge or separate the object boundary blocks, and extracts the object boundary block merging / separation information accordingly. An extraction unit, The object boundary block is input by using the image information padded by the out-of-object region padding unit, and the reconstructed shape information input from the shape information encoder and the merge / separation information input from the object boundary block merge / separation information extractor. An object boundary block merging unit for merging image information of A DCT unit which receives the image information from the object boundary block merging unit and performs DCT; A quantization unit configured to receive and quantize the DCT coefficients from the DCT unit; A DCT coefficient and a related information encoder for variable length encoding by receiving the quantized signal from the quantization unit and using the merge / separation information input from the object boundary block merge / separation information extractor; An inverse quantizer configured to receive and inverse quantize the quantized signal from the quantizer; An inverse DCT unit configured to receive an inverse quantized signal from the inverse quantizer and perform inverse DCT to output a decoded signal; A merged block separator which receives the signal decoded by the inverse DCT unit and outputs reconstructed image information by separating the merged block using merge / separation information input from the object boundary block merge / separation information extractor; , Receives shape information encoded by the shape information encoder, receives motion information estimated by the motion estimation unit, receives image information encoded by the DCT coefficient and the related information encoder, multiplexes a bitstream through a buffer MPEG-4 encoder, characterized in that consisting of a multiplexer for transmitting. The method of claim 4, wherein An adder for receiving the video information of the motion-compensated VOP from the motion compensator and receiving the reconstructed video information from the merged block separator and adding the two pieces of information; MPEG-4 encoder, characterized in that the addition of the VOP detection unit for receiving the addition value from the adder detects the image information of the previous VOP and inputs to the motion estimation and motion compensation to use the motion estimation and motion compensation . In the method of encoding the original image information or image error information using shape information, A first step of padding an area outside the object by using shape information for padding original image information or image error information; A second step of merging the image information of the object boundary block using the merged / separated information extracted from the padded image information in the first step and the restored shape information; A third step of performing DCT and quantizing the image information merged in the second step; And a fourth step of variable length encoding the extracted quantized DCT coefficients using the merged / separated information extracted in the third step. The method of claim 6, A fifth step of converting the decoded signal by performing inverse quantization and inverse DCT on the signal quantized in the third step; And a sixth step of reconstructing image information by separating the merged block by using the merged / separated information from the signal decoded in the fifth step. The method of claim 6, In the method of extracting merge / separation information in the second step, when the reconstructed shape information, which is the basis of extraction, is a macroblock composed of four 8 × 8 blocks, the horizontal merge / information extraction process is performed by the upper and left upper blocks. If there is no overlapping pixel among objects in the object between the shape information of the upper right block 180 degrees and the shape information of the upper left block, the image information of the upper left block is rotated 180 degrees. Extracting a flag to treat the upper right block as an out-of-object block when encoding image information; If the lower left block and the lower right block are object boundary blocks, and the shape information of the lower right block is rotated 180 degrees and there are no pixels overlapping each other among the pixels in the object between the shape information of the lower left block, the image information of the lower right block is displayed. The image information encoding method of the MPEG-4 encoder, comprising: extracting a flag to be rotated 180 degrees and merged into image information of the lower left block, and the right lower block is treated as an out-of-object block when encoding image information. The method of claim 6, In the method of extracting merge / separation information in the second step, when the reconstructed shape information, which is the basis of extraction, is a macroblock composed of four 8 × 8 blocks, the vertical merge / separation information extraction process is performed on the upper left and lower blocks. Block, the shape information of the lower left block is rotated 180 degrees and if there is no overlapping pixel among the pixels in the object between the shape information of the upper left block, the image information of the upper left block is rotated 180 degrees. Extracting a flag to be merged with the information and to treat the left lower block as an out-of-object block when encoding image information; If the upper and lower blocks are object boundary blocks, and the shape information of the lower right block is rotated 180 degrees and there is no pixel overlapping among the pixels in the object between the shape information of the upper right block, the image information of the lower right block is 180 degrees. And rotating to merge the image information of the upper right block, and extracting a flag to treat the right lower block as an out-of-object block when encoding the image information. The method of claim 6, In the method of extracting merge / separation information in the second step, when the reconstructed shape information, which is the basis of extraction, is a macroblock composed of four 8 × 8 blocks, the diagonal merge / separation information extraction process includes the upper left and lower right blocks. If it is an object boundary block and the shape information of the lower right block is rotated 180 degrees and there is no overlapping pixel among the pixels in the object between the shape information of the upper left block, the image information of the lower right block is rotated 180 degrees and the upper left block is Extracting a flag to be merged into the image information of and to treat the lower right block as an out-of-object block when encoding the image information; If the upper right block and the lower left block are object boundary blocks, and the shape information of the lower left block is rotated by 180 degrees and there is no pixel overlapping with each other among the pixels in the object between the shape information of the upper right block, the image information of the lower left block is displayed. The image information encoding method of the MPEG-4 encoder, comprising: extracting a flag to be rotated by 180 degrees and merged into the image information of the upper right block, and the lower left block to be treated like an out-of-object block when encoding the image information. The method of claim 6, The method of extracting merge / separation information in the second step may include: a first method of extracting the horizontal merge / separation information, the vertical merge / separation information, and then extracting the diagonal merge / separation information; A second method of extracting the horizontal merge / separation information, extracting the diagonal merge / separation information, and then extracting the vertical merge / separation information; A third method of extracting the vertical merge / separation information, extracting the horizontal merge / separation information, and then extracting the diagonal merge / separation information; A fourth method of extracting the vertical merge / separation information, extracting the diagonal merge / separation information, and then extracting the horizontal merge / separation information; A fifth method of extracting the diagonal merge / separation information, the horizontal merge / separation information, and then extracting the vertical merge / separation information; And a sixth method of extracting the diagonal merger / separation information, the vertical merger / separation information, and then extracting the horizontal merger / separation information. The method of claim 11, The first, second, third, fourth, fifth, and sixth merge / separation information extraction methods may be omitted if a flag occurs in any one of the three merge / separation information extraction processes. Image information encoding method of MPEG-4 encoder. The method according to any one of claims 8 to 12, When the image format of the image information input for encoding includes only luminance blocks without color difference blocks or consists of four blocks such as transparency information of gray scale shape information, the merge / separation information extraction process may be performed by any of the first to sixth methods. The video information encoding method of the MPEG-4 encoder, characterized in that one method. The method according to any one of claims 8 to 12, The merge / separation information extraction process in the case where the image format of the image information input for encoding is 4 (Y): 2 (Cb): 0 (Cr) is performed by using any one of the first to sixth methods only for the luminance block. Image information encoding method of the MPEG-4 encoder characterized in that applied. The method according to any one of claims 8 to 12, The merge / separation information extraction process in the case where the image format of the image information input for encoding is 4 (Y): 2 (Cb): 2 (Cr) may be any of the first to sixth methods for the luminance block. Apply, The image information encoding method of the MPEG-4 encoder, characterized in that for the chrominance block, only the horizontal merge / separation information extraction process is applied for each color difference component (Cb, Cr). The method according to any one of claims 8 to 12, The merge / separation information extraction process in the case where the image format of the image information input for encoding is 4 (Y): 4 (Cb): 4 (Cr) may be any of the first to sixth methods for the luminance block. Apply, The image information encoding method of the MPEG-4 encoder, characterized in that for the chrominance block, any one of the first to sixth methods is applied for each chrominance component (Cb, Cr). The method of claim 6, The method of merging the image information of the object boundary block by receiving the padded image information in the second step and using the extracted merge / separation information and the reconstructed shape information includes luminance of the padded macroblock unit. An object boundary block merging step of the luminance block applying the object boundary block merger to the block; A video format determination step of judging whether a color difference block is present in the input image information and whether or not the video format is 4: 2: 0; In the image format determination step, when the format of the input image is 4: 2: 2 or 4: 4: 4, the object boundary block merges the chrominance block to apply the object boundary block merger to the chrominance block by color difference component (Cb, Cr). Image information encoding method of the MPEG-4 encoder, characterized in that the application step. The method of claim 17, And the chrominance block is terminated when the color format block does not exist or the image format is 4: 2: 0. The method of claim 17, In the object boundary block merging step of the luminance block, a block in which merge / separation information of any one of the four 8 × 8 blocks in the 16 × 16 luminance block Bi is received and the block Bi merges the other blocks. Determining whether or not the merged block to determine whether the recognition; A block merge step of merging the block Bi and a block merged with the block if the block Bi merges in the merge block determination step; I number increasing step of increasing the number of i by one, I number determining step of determining whether i is less than 4; And if i is smaller than 4 in the i number determination step, proceeding to the merge block determination step, and ending with i when the i number is 4, the video information encoding method of the MPEG-4 encoder. The method of claim 19, And if the block (Bi) is not a block for merging other blocks in the step of determining whether the merged block is present, the i number increasing step is performed. The method of claim 19, In the block merging step, the merged block is rotated 180 degrees and then merged into the merged block. At this time, the image information of the object area of the merged block rotated 180 degrees replaces the padded information of the merged block, while the 180 is merged. And the padded information of the out-of-object region of the merged block to be rotated by an average value with the padded information of the merged block. The method of claim 17, In the object boundary block merging step of the chrominance block, when the input image format is 4: 2: 2, the merge / separation information of one block (Bi) of two 8 × 8 blocks in the chrominance information (Cb or Cr) is received. A judging step of judging whether or not the block Bi is a block of merging other blocks; A block merging step of merging the block Bi and a block merged with the block if the block Bi merges in the merging block determination step;   I number increasing step of increasing the number of i by one, I number determining step of determining whether i is less than 2; And if i is less than 2 in the i number determination step, proceeding to the merged block determination step, and terminating when i becomes 2, the method of encoding image information of an MPEG-4 encoder. The method of claim 22, And if the block (Bi) is not a block for merging other blocks in the merge block determination step, proceeding to increasing the number of i, the image information encoding method of the MPEG-4 encoder. The method of claim 22, In the block merging step, the merged block is rotated 180 degrees and then merged into the merged block. At this time, the image information of the region within the object of the merged block rotated 180 degrees replaces the padded information of the merged block. The padding information of the object outside of the merged block rotated by 180 degrees is replaced with the averaged value with the padded information of the merged block. The method of claim 17, In the object boundary block merging step of the color difference block, when the input image format is 4: 4: 4, merge / separation information of one block Bi among two 8 × 8 blocks in the color difference information Cb or Cr is received. A judging step of judging whether or not the block Bi is a block of merging other blocks; A block merging step of merging the block Bi and a block merged with the block if the block Bi merges in the merging block determination step;   I number increasing step of increasing the number of i by one, I number determining step of determining whether i is less than 4; And if i is smaller than 4 in the i number determination step, proceeding to the merged block determination step, and ending with i when the i is 4, the video information encoding method of the MPEG-4 encoder. The method of claim 25, And if the block (Bi) is not a block for merging other blocks in the merge block determination step, proceeding to increasing the number of i, the image information encoding method of the MPEG-4 encoder. The method of claim 7, wherein The method for separating the merged block by using the merged / separated information by receiving the signal decoded in the sixth step includes the object boundary of the luminance block that applies the object boundary block separation to the luminance block among the image information received in the macroblock unit. Block separation application step, A video format determination step of judging whether a color difference block exists in the input image information and whether or not the video format is not 4: 2: 0; In the image format determination step, when the format of the input image information is 4: 2: 2 or 4: 4: 4, the object boundary block of the chrominance block that applies object boundary block separation to the chrominance block by color difference components (Cb, Cr). The video information encoding method of the MPEG-4 encoder, characterized in that the step of applying a separation. The method of claim 27, The image information encoding method of the MPEG-4 encoder, characterized in that in the image format determination step, if the chrominance block does not exist or the image format is 4: 2: 0. The method of claim 27, In the object boundary block separating and applying step of the luminance block, the merge / separation information of one of the four 8 × 8 blocks in the 16 × 16 luminance block is input, and whether the block Bi merges the other blocks. Determining whether or not the merger with other blocks to determine the; A block separation step of separating a merged block from the block Bi if the block Bi merges another block in the determination of whether to merge the other blocks; I number increasing step of increasing the number of i by one, An i number determining step of determining whether i is less than 4 in the increasing step; And if i is less than 4 in the i number determination step, proceeding to the other block merging determination step, and ending with i when the i number is 4, the video information encoding method of the MPEG-4 encoder. The method of claim 29, And if the block (Bi) is not a block in which another block is merged in the step of determining whether to merge with other blocks, the i number increasing step is performed. The method of claim 29, And the block separating step replaces all the image information of the merged block with the image information rotated 180 degrees of the merged block. The method of claim 27, In the object boundary block separation step of the chrominance block, when the input image format is 4: 2: 2, the merge / separation information of one block (Bi) of two 8 × 8 blocks in the chrominance information (Cb or Cr) is received. Determining whether or not the other block is merged with the block Bi; A block separation step of separating a merged block from the block Bi if the block Bi merges another block in the determination of whether to merge the other blocks; I number increasing step of increasing the number of i by one, I number determining step of determining whether i is less than 2; In the i number determination step, if i is less than 2, the process proceeds to the other block merger determination step. On the other hand, when i becomes 2, the video information encoding method of the MPEG-4 coder is finished. . The method of claim 32, And if the block (Bi) is not a block in which another block is merged in the step of determining whether to merge with other blocks, the i number increasing step is performed. The method of claim 32, And the block separating step replaces all the image information of the merged block with the image information rotated 180 degrees of the merged block. The method of claim 27, In the object boundary block separating and applying step of the color difference block, when the input image format is 4: 4: 4, merge / separation information of one block (Bi) among four 8 × 8 blocks in the color difference information (Cb or Cr) is received. Determining whether or not the other block is merged with the block Bi; A block separation step of separating a merged block from the block Bi if the block Bi merges another block in the determination of whether to merge the other blocks; I number increasing step of increasing the number of i by one, I number determining step of determining whether i is less than 4; In the i number determination step, if i is less than 4, the process proceeds to the other block merger determination step. On the other hand, when i becomes 4, the video information encoding method of the MPEG-4 encoder comprises an end step. . The method of claim 35, wherein If the block (Bi) is not a block to merge the other block in the step of determining whether or not the merge of the other blocks, the i-information encoding method of the MPEG-4 encoder, characterized in that proceeds to the increasing step. An image information encoding apparatus for receiving original image information or image error information and encoding the same using shape information. An object boundary block merging / separation information extracting unit for determining whether or not to merge the object boundary block and receiving the reconstructed shape information and extracting the object boundary block merging / separation information accordingly; An object boundary block merging unit for merging image information of an object boundary block by using shape information restored by receiving original image information or image error information and merging / separation information input from the object boundary block merging / separation information extracting unit; ,   Shape adaptation by performing image adaptation DCT using the merged / separated information input from the object bound block merger / separation information extractor and the shape information of the merged image by receiving the image information merged by the object boundary block merger With type DCT part, A quantization unit configured to receive and quantize the shape-adaptive DCT coefficients from the shape-adaptive DCT unit; Using the merged / separated information input from the object boundary block merged / separated information extractor and receiving shape information of the merged image, the quantized shape-adaptive DCT coefficient is inputted by the quantization unit, and skips the region where the coefficient does not exist. And a shape-adaptive DCT coefficient and an associated information encoder for performing an adaptive scan applying a zigzag scan. The method of claim 37, wherein An inverse quantizer configured to receive and inverse quantize the quantized signal from the quantizer; Inversely adaptive DCT that performs inversely adaptive DCT using the merger / separation information and the merged shape information input from the object boundary block merger / separation information extraction unit by receiving the dequantized signal from the inverse quantization unit. Wealth, The merged block separation unit for receiving the signal decoded by the inverse-adaptive DCT unit and reconstructing the image information by separating the merged blocks using the merge / separation information input from the object boundary block merge / separation information extractor The video information encoding apparatus of the MPEG-4 encoder, characterized in that the addition is made. A VOP forming unit for forming a VOP, meaning a content object having an arbitrary shape that can be accessed and edited by a user at a point in time on the time axis; A shape information encoder which receives the shape information of the VOP from the VOP forming unit, encodes, decodes and outputs the original image shape information which is currently input; The image information of the VOP formed by the VOP forming unit is received, the image information of the previous VOP reproduced is received, and the shape information encoder receives the shape information or original image shape information reconstructed by the shape information encoder to estimate the motion in macroblock units. A motion estimation unit for outputting motion information, A motion compensator for receiving motion information from the motion estimation unit, receiving shape information reconstructed by the shape information encoder, and outputting image information of the motion-compensated VOP by receiving the reproduced image information of the previous VOP; , A subtractor for receiving the image information of the VOP whose motion is compensated by the motion compensator and obtaining the difference between the two image information by receiving the image information of the VOP formed by the VOP forming unit; Object boundary block merging / separation information extraction for determining whether or not to merge object boundary blocks by receiving shape information restored by the shape information encoder and extracting object boundary block merging / separation information accordingly Wealth, The subtractor receives the original image information or the image error information by using the reconstructed shape information input from the shape information encoder and the merge / separation information input from the object boundary block merging / separation information extractor. An object boundary block merging unit for merging image information,   Shape adaptation by performing image adaptation DCT using the merged / separated information input from the object bound block merger / separation information extractor and the shape information of the merged image by receiving the image information merged by the object boundary block merger With type DCT part, A quantization unit configured to receive and quantize the shape-adaptive DCT coefficients from the shape-adaptive DCT unit; Using the merged / separated information input from the object boundary block merged / separated information extractor and receiving shape information of the merged image, the quantized shape-adaptive DCT coefficient is inputted by the quantization unit, and skips the region where the coefficient does not exist. A shape-adaptive DCT coefficient and related information encoder for performing an adaptive scan applying a zigzag scan An inverse quantizer configured to receive and inverse quantize the quantized signal from the quantizer; Inversely adaptive DCT that performs inversely adaptive DCT using the merger / separation information and the merged shape information input from the object boundary block merger / separation information extraction unit by receiving the dequantized signal from the inverse quantization unit. Wealth, A merged block separator that receives the signal decoded by the inversely adaptive DCT unit and reconstructs image information by using the merged / separated information input from the object boundary block merger / separation information extractor to reconstruct image information. Wow, Receiving shape information encoded by the shape information encoder, receiving motion information estimated by the motion estimation unit, and receiving multiplexed image information encoded by the shape-adaptive DCT coefficient and the related information encoder. MPEG-4 encoder, characterized in that consisting of a multiplexer for transmitting a bitstream through. The method of claim 39, An adder for receiving the video information of the motion-compensated VOP from the motion compensator and receiving the reconstructed video information from the merged block separator and adding the two pieces of information; MPEG-4 encoder, characterized in that the addition of the VOP detection unit for receiving the addition value from the adder detects the image information of the previous VOP and inputs to the motion estimation and motion compensation to use the motion estimation and motion compensation . In the method of encoding the original image information or image error information using shape information, A first step of determining whether to merge or separate the object boundary blocks using the restored shape information, and extracting the object boundary block merging / separation information accordingly; A second step of merging the image information of the object boundary block by using the shape information reconstructed from the original image information or the image error information and the merge / separation information extracted in the first step; A third step of performing shape adaptation DCT by receiving the merged / separated information and the shape information of the merged image from the image information merged in the second step; A fourth step of quantizing the shape-adaptive DCT coefficient obtained in the third step; An adaptive scan is performed by applying a zigzag scan by skipping the region having no coefficients using the merged / separated information and the shape information of the merged image by receiving the quantized shape-adaptive DCT coefficients obtained in the fourth step. And a fifth step of encoding the encoded video information. The video information encoding method of the MPEG-4 encoder. 42. The method of claim 41 wherein A sixth step of inverse quantization of the quantized signal obtained in the fourth step; A seventh step of performing inverse-adaptive DCT on the inverse quantized signal obtained in the sixth step by using the merged / separated information and the merged shape information; And an eighth step of reconstructing image information by dividing the merged block by using the merged / separated information from the decoded signal obtained in the seventh step. 42. The method of claim 41 wherein In the second step of merging the image information of the object boundary block, the image information of the merged block is rotated 180 degrees and then merged into the merged block. In this case, the image information of the intra-object region of the merged block rotated 180 degrees. Replaces information of the out-of-object area of the merged block and DON'T CARE for all information of the out-of-object area merged with the merged block. In the encoding apparatus using the shape information by receiving the original image information or image error information, An object boundary block merging / separation information extracting unit for determining whether or not to merge the object boundary block and receiving the reconstructed shape information and extracting the object boundary block merging / separation information accordingly; An object boundary block merging unit for merging image information of an object boundary block by using shape information restored by receiving original image information or image error information and merging / separation information input from the object boundary block merging / separation information extracting unit; , Arbitrary / separation information is inputted from the object boundary block merging / separation information extracting unit, and arbitrary shape generated by using shape information (original image shape information, restored shape information or original image shape information and restored shape information) for padding A padding shape information extracting unit for receiving shape information, etc., and outputting shape information for padding of the merged and unmerging blocks of the object boundary block; An out-of-object area padding unit which receives image information from the object boundary block merging unit and pads the out-of-object area using shape information for padding the merged block and the unmerged block inputted by the padding shape information extracting unit; , A DCT unit configured to receive image information from the area padding unit outside the object and perform DCT; A quantization unit configured to receive and quantize the DCT coefficients from the DCT unit; MPEG-4, characterized in that consisting of the DCT coefficient and the associated information encoding unit for variable-length encoding using the merge / separation information input from the object boundary block merge / separation information extraction unit receives the quantized DCT coefficient from the quantization unit Image information encoding apparatus of the encoder. The method of claim 44, An inverse quantizer configured to receive and inverse quantize the quantized signal from the quantizer; An inverse DCT unit configured to receive an inverse quantized signal from the inverse quantizer and perform inverse DCT; A merged block separator which receives the signal decoded by the inverse DCT unit and separates the merged block using the merge / separation information input from the object boundary block merge / separation information extractor to reconstruct image information The video information encoding apparatus of the MPEG-4 encoder. A VOP forming unit for forming a VOP, meaning a content object having an arbitrary shape that can be accessed and edited by a user at a point in time on the time axis; A shape information encoder which receives the shape information of the VOP from the VOP forming unit, encodes, decodes and outputs the original image shape information which is currently input; The image information of the VOP formed by the VOP forming unit is received, the image information of the previous VOP reproduced is received, and the shape information encoder receives the shape information or original image shape information reconstructed by the shape information encoder to estimate the motion in macroblock units. A motion estimation unit for outputting motion information, A motion compensator for receiving motion information from the motion estimation unit, receiving shape information reconstructed by the shape information encoder, and outputting image information of the motion-compensated VOP by receiving the reproduced image information of the previous VOP; , A subtractor for receiving the image information of the VOP whose motion is compensated by the motion compensator and obtaining the difference between the two image information by receiving the image information of the VOP formed by the VOP forming unit; Object boundary block merging / separation information which receives the shape information restored by the shape information encoder and determines whether to merge or separate object boundary blocks and extracts object boundary block merging / separation information accordingly. An extraction unit, The subtractor receives the original image information or the image error information by using the restored shape information inputted from the shape information encoder and merged / separated information inputted from the object boundary block merger / separation information extractor. An object boundary block merging unit for merging image information, The merge / separation information is input from the object boundary block merging / separation information extracting unit, and the shape information encoding unit receives shape information (original image shape information, restored shape information, or original image shape information and restored shape information) for padding. A padding shape information extracting unit for receiving shape information, etc. generated using the same, and outputting shape information for padding of the merged and unmerged blocks of the object boundary block; An out-of-object area padding unit which receives image information from the object boundary block merging unit and pads the out-of-object area using shape information for padding the merged block and the unmerged block inputted by the padding shape information extracting unit; , A DCT unit configured to receive image information from the area padding unit outside the object and perform DCT; A quantization unit configured to receive and quantize the DCT coefficients from the DCT unit; A DCT coefficient and a related information encoder for variable length encoding by receiving the DCT coefficient quantized by the quantization unit using merge / separation information input from the object boundary block merge / separation information extracting unit; An inverse quantizer configured to receive and inverse quantize the quantized signal from the quantizer; An inverse DCT unit configured to receive an inverse quantized signal from the inverse quantizer and perform inverse DCT; A merged block separation unit configured to receive a signal decoded by the inverse DCT unit and to reconstruct image information by separating the merged block using merge / separation information input from the object boundary block merge / separation information extracting unit; Receives shape information encoded by the shape information encoder, receives motion information estimated by the motion estimation unit, receives image information encoded by the DCT coefficient and the related information encoder, multiplexes a bitstream through a buffer MPEG-4 encoder, characterized in that consisting of a multiplexer for transmitting. The method of claim 46, An adder for receiving the video information of the motion-compensated VOP from the motion compensator and receiving the reconstructed video information from the merged block separator and adding the two pieces of information; MPEG-4 encoder, characterized in that the addition of the VOP detection unit for receiving the addition value from the adder detects the image information of the previous VOP and inputs to the motion estimation and motion compensation to use the motion estimation and motion compensation . In the method of encoding the original image information or image error information using shape information, A first step of judging whether to merge or separate the object boundary block as the restored shape information and extracting the object boundary block merging / separation information accordingly; A second step of merging the image information of the object boundary block by using the shape information reconstructed from the original image information or the image error information and the merge / separation information extracted in the first step; By using the merge / separation information extracted in the first step and shape information (pad shape information, restored shape information or arbitrary shape information generated using the original image shape information and restored shape information) for padding A third step of obtaining shape information for padding of the merged and unmerged blocks of the object boundary block; A fourth step of padding the out-of-object region by using the merged image information obtained in the second step using shape information for padding the merged block and the unmerged block obtained in the third step; A fifth step of performing DCT on the image information padded in the fourth step; A sixth step of quantizing the DCT coefficient obtained in the fifth step; And a seventh step of variable length encoding the quantized DCT coefficient obtained in the sixth step using the merge / separation information extracted in the first step. The method of claim 48, An eighth step of inverse quantizing the quantized signal obtained in the sixth step, A ninth step of performing inverse DCT on the inverse quantized signal obtained in the eighth step; The video information encoding of the MPEG-4 encoder is performed by receiving the decoded signal obtained in the ninth step and reconstructing the image information by separating the merged block using the merge / separation information. Way.