CN1791218A - Variable length coding method and variable length decoding method. - Google Patents

Abstract

A method for segmentally encoding and decoding objects in an image, the encoding is to obtain segment data in which the object is divided into segments, the segment data corresponds to a plurality of data segments of a plurality of objects, and is suitable for The pattern determined by the data distribution characteristics of this section scans and codes the obtained data, and outputs the coded data and the scanning pattern data. The decoding is to receive the coded data for decoding, and arrange the decoded data according to the position determined by the scanning pattern data to reconstruct a plurality of two-dimensional data. The encoding and decoding of the present invention can adopt variable length method, Huffman method or arithmetic method.

Description

Method for encoding and decoding objects in an image by segment data

This application is a divisional application for an invention patent application with the filing date of January 24, 1997, the title "Coding and Decoding Method for Segment Data of Objects in Images", and the application number 00126057.x.

technical field

The present invention relates to a digital data encoding and decoding system, in particular to a method for encoding and decoding a target object in an image with segmented digital data, so as to further improve the compression rate of stored or transmitted data.

Background technique

Recently, in systems for transmitting and receiving video and audio signals, there have been multiple methods of encoding video and audio signals transmitted or stored in storage media into digital data and decoding the encoded digital data for reproducing video and audio signals. way. However, people have searched for a technical method to further compress the amount of transmitted or stored data so as to improve the efficiency of data transmission in the coding and decoding system. Examples of encoding methods for transmitted or stored digital data include transform encoding, differential pulse code modulation (DPCM), vector quantization, and variable length encoding. These encoding methods further compress the overall data volume by removing redundant data from transmitted or stored digital data.

The above-mentioned existing compression methods all use a square block of a certain size such as 8×8 or 16×16 as a basic unit for data processing. However, the image data processing method involved in the present invention is not limited to all image frames in units of blocks, but only processes meaningful objects in the image by segment. That is, video data of each frame is divided into segments of a set length, and data processing is performed in an encoding and decoding system for storage, transmission, and reception of video signals. Each segment data or differential data between segment data is orthogonally transformed, and the video data is transformed into transform coefficients in the frequency domain. Well-known segment data transformation methods include discrete cosine transform (DCT), Walsh Hadamard transform (WHT), discrete Fourier transform (DFT), and discrete sine transform (DST). Transform coefficients obtained by these transform methods are appropriately coded according to the characteristics of coefficient data, thereby improving compression efficiency. Since human vision is more sensitive to low frequencies than to high frequencies, high frequency data is reduced by data processing. Thus, the amount of encoded data can be reduced.

Contents of the invention

The purpose of the present invention is to provide an encoding method for the prior art. This encoding method encodes segmented data of a target object within an image using a scanning pattern most suitable for distribution characteristics of each segment data, and the segment data of the object corresponds to a plurality of data segments of a plurality of objects.

Another object of the present invention is to provide a decoding method that uses the same scanning pattern as that selected in the encoding process of each segment data to perform arithmetic decoding on data encoded by segment data of a target object in an image. code, and the segment data of the object corresponds to a plurality of data segments.

In order to achieve the above object, the present invention provides a method for encoding digital data in which an object in an image is divided into segments. The method includes the following steps: obtaining segment data of the object, the segment data being A plurality of data segments of an object; obtaining its transformation coefficient for the segment data of the above-mentioned object; scanning the data obtained in the above-mentioned steps according to the selected scanning mode suitable for the segment data distribution characteristics of the object, And encode the scanned data; output the above-mentioned encoded data and the scanning mode data of the scanning mode adopted when encoding the data, wherein the segment data represents the pixel value selected by the object in the image and related to the object pixel value The size and shape of the corresponding image segment.

In the present invention as described above, not only the scanned and coded data correspond to a plurality of segment data of a plurality of objects in the encoding step, but also the scanning range of the scanning mode selected by the distribution characteristics of the segment data covers only the segment. The distribution area of the data, in order to more effectively encode the image digital data.

The data obtained in the above-mentioned step of obtaining the transformation coefficient thereof is a set coefficient and corresponds to a plurality of two-dimensional data conforming to the distribution characteristics of the segment data of the object.

In the scanning mode selected according to the distribution characteristics of the segment data, the scanning direction can be horizontal or vertical, or a scanning direction with an inclination of 30 degrees or 45 degrees.

In order to select the scanning mode most suitable for the distribution characteristics of the segment data as described above, in the above encoding step, the obtained data are scanned and encoded respectively according to the given multiple scanning modes, and the data corresponding to each encoded data are selected. The scan mode corresponding to the minimum of the accumulated length values.

The scanning mode selected in this way not only has the flexibility to change with the distribution characteristics of the segment data, but also is most suitable for the scanning of the segment data. According to the requirements of digital data processing, variable length coding or Huffman coding can be used for the scanned data in the above coding step, and arithmetic coding can also be used.

To achieve another object of the present invention, the present invention provides a method for decoding encoded data, the method comprising the following steps: receiving the above-mentioned encoded data for decoding; The provided scan pattern data is arranged according to the location, wherein the received encoded data corresponds to a plurality of data segments, the encoded data being data encoding objects within the image segment by segment.

In the decoding method described above, the received encoded data can be multiple data segments corresponding to one image, or multiple data segments corresponding to multiple images; and the scan mode data is Supplied with encoded data and represents the scan information used for that encoded data. That is, it indicates the scanning mode adopted to suit the data distribution characteristics of the target segment and the information that scans are performed only within the data distribution range of the segment. Therefore, in the arranging step above, the decoded data is arranged according to the scanning direction and scanning range specified by the scanning pattern data.

Furthermore, a scan address is generated corresponding to the scan direction and scan range determined according to the provided scan pattern data, and the decoded data are arranged one by one at the positions specified by the scan address. The data arranged in the above steps actually constitute a plurality of two-dimensional data or a plurality of two-dimensional set coefficients, and the distribution characteristics of the two-dimensional data or two-dimensional set coefficients are the same as those of the object before encoding The distribution characteristics of each segment of data are consistent.

Therefore, the scanning direction of the scanning address generated according to the scanning pattern data or the provided scanning pattern can be a horizontal direction or a vertical direction, or a scanning direction with an inclination angle of 30 degrees or 45 degrees. Furthermore, in the decoding step, variable length decoding, Huffman decoding or arithmetic decoding is adopted correspondingly according to the type of the received encoded data.

The present invention may be used for the purpose of encoding and/or decoding data transmitted or stored in semiconductor memory, magnetic tape, magnetic disk, CD-ROM, digital video disk or other storage media and data retrieved therefrom.

Description of drawings

Figure 1 is a block diagram of an embodiment of a known encoder using a variable length encoder.

Figure 2 is a block diagram of an embodiment of a conventional decoder using a variable length decoder.

3A to 3C are diagrams illustrating examples of conventional methods of digital data segmentation, scanning patterns, and encoding processing.

3D to 3G are exemplary diagrams illustrating the present invention according to digital data segmentation, scanning patterns, and encoding processes of various objects.

Fig. 3H is a flow chart of POCS-based arbitrary morphological transformation.

Fig. 4 is a block diagram illustrating the distribution state of variable-length coded data.

Fig. 5 is a block diagram showing an embodiment of a variable length coder according to the present invention.

Fig. 6 is a block diagram showing an embodiment of a variable length decoder according to the present invention.

7A to 7C are diagrams illustrating scanning patterns used in connection with FIGS. 5 and 6 .

8A to 8C are diagrams showing simple examples of multi-segment scanning according to the present invention.

8D to 8E are diagrams showing simple examples of multi-segment scanning patterns according to the present invention.

Detailed ways

Below, the preferred embodiment of the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 is a block diagram schematically showing a conventional encoding apparatus for video data using a variable length encoding method. The input 10 receives data segments corresponding to block-wise segmented image segments or data segments corresponding to object-wise segmented image segments. Generally, the size of M1×M2 is used to represent the block segment, but for the sake of illustration, it is assumed that M1=M2=M3. The segment data input through the input terminal 10 is added to the set feedback data in the first adder A1 to calculate the difference data between two sets of data (ie input data and feedback data). The orthogonal transformer 11 performs discrete cosine transform on the input difference data, and converts the difference data into coefficients in the frequency domain. The quantizer 12 changes the transform coefficients into multi-level representative values through the set quantization process. At this time, the quantizer 12 variably quantizes the data output from the orthogonal transformer 11 according to the quantization level Q input from the buffer 14 . The variable length encoder 13 performs variable length encoding on the segment data according to the statistical properties of the quantized coefficients and generates compressed data V _CD . The process of variable-length coding video data will be described later. The buffer 14 receives the compressed data from the variable length coder 13, and outputs the data to the transmission channel at a fixed rate. At this time, the quantization level Q is output, and its purpose is to control the amount of compressed data in order to prevent overflow and underflow of data.

In general, similar patterns exist between adjacent frames of video data. Thus, when the image moves slightly, the motion of the image can be inferred from the comparison of the current frame and the past frame. The calculation of the motion vector MV relies on the result of motion inference, and the motion compensation can be obtained from the past frame according to the motion vector. Since the amount of differential data between the segment data obtained from motion compensation and the segment data input to the input terminal 10 is very small, the data can be further compressed by the encoding process described above. A feedback loop for motion estimation and motion compensation is composed of an inverse quantizer 15 , an inverse orthogonal transformer 16 , a frame memory 17 , a motion estimator 18 and a motion compensator 19 . The inverse quantizer 15 inversely quantizes the quantized coefficients output from the quantizer 12, and the inverse orthogonal transformer 16 converts the output data of the inverse quantizer 15 into video data in the spatial domain through inverse discrete cosine transform. The second adder A2 outputs segment data resulting from adding the video data output from the inverse orthogonal converter 16 to the feedback data input from the second switch SW2. The segment data output from the second adder A2 is sequentially stored in the frame memory 17, thereby reconstructing a frame. The motion inferring unit 18 finds the segment data of the most similar pattern to the segment data input through the input terminal 10 from the frame data stored in the frame memory 17, and calculates the motion vector MV used for image motion estimation from these two segment data. . In order to be able to use the motion vector MV in the decoding system, it is transmitted to the receiver and motion compensator 19 . The motion compensator 19 reads segment data corresponding to the motion vector MV from the frame data in the frame memory 17, and supplies the read data to the first adder A1. As previously described, the first adder A1 calculates differential data between the segment data input from the input terminal 10 and the segment data input from the motion compensator 19, encodes the differential data, and transmits it to the receiver. Furthermore, the two switches SW1 and SW2 in Fig. 1 are update switches, the purpose of which is to prevent the difference between the encoded frame data and the unprocessed frame data from being generated by the accumulation of differential data, and update the data to the set A frame or segment unit of a given size.

The coded video data V _CD is either stored in a storage medium or transmitted in a receiver to an input to a decoder as shown in FIG. 2 . The variable length decoder 21 decodes the input video data V _CD through the inverse process of variable length coding. The inverse quantizer 22 decodes the quantized coefficients input from the variable length decoder 21, and outputs transform coefficients in the frequency domain. The inverse orthogonal transformer 23 transforms the transform coefficients in the frequency domain input from the quantizer 22 into video data in the spatial domain. The motion vector MV output from the motion estimator 18 of the encoder is input to the motion compensator 24 of the decoder. The motion compensator 24 reads out the segment data corresponding to the motion vector from the frame data stored in the frame memory 25, and adds the read data to the adder A. The adder A adds the differential data output from the inverse forward commutator 23 to the segment data input from the motion compensator 24, and outputs reconstructed segment data resulting from the result. A switch SW is connected to the output of the motion compensator and acts in the same way as the refresh switch in the encoder of Fig. 1 described above.

In order to achieve the purpose of variable length coding, the existing coding system uses Huffman (Huffman) coding method. The Huffman coding method is to assign codes of different lengths to the input data according to the probability of the set symbols. That is to say, the higher the probability is, the shorter the code is assigned, and the lower the probability is, the longer the code is assigned. In encoding using the Huffman algorithm, there are a large number of different symbols. When a specific symbol has a low probability, when a long code is assigned to a large number of symbols that rarely appear according to the Huffman algorithm, the encoding and Data processing becomes more complicated in the decoding process. In order to solve this problem, if a set fixed-length code can be assigned to the distribution field of many symbols that rarely appear (hereinafter referred to as the escape field), even if the average code length is higher than the average value of the Huffman code, the data Processing complexity can also be greatly reduced.

Fig. 3A shows an example of the data structure divided into segments of 8x8 size; Fig. 3B shows the transformation of 8x8 segment data to the frequency domain and the 8x8 quantization coefficients generated according to the quantized transformed data; Fig. 3C shows that in view of A large number of quantized coefficients are "0" in the frequency domain, Z-shaped scanning of quantized coefficients from low frequency to high frequency and encoding of scanned coefficients to [length, level] symbols. In the run-level coding method, "run" means the number of "0" occurrences among non-"0" coefficients, and "level" means the absolute value of non-"0" coefficients. In the case of 8×8 data in Figure 3A to Figure 3C, the value of "run" is from "0" to "63", when the quantized output has an integer value from "-255" to "255", "level" The value is from "1" to "255". Its code indicates otherwise.

FIG. 3D shows another method of segmenting image data in a manner consistent with another object of diverse segmentation. In general, each data segment corresponds to a separately shaped object. Fig. 3E shows an example of encoding of data segments corresponding to objects. Since the shape of the object is arbitrary, a special transformation method of traditional block-based orthogonal transformation on a rectangle surrounding a given object is necessary. A given object can best be reconstructed according to the selected number of transform coefficients by proper selection of pixel values outside the given object and within the rectangle surrounding it. FIG. 3F shows an example of selecting transform coefficients. In the case of a data segment representing an image portion having L pixels, there may be L meaningful transform coefficients in the transform domain. Depending on the selection of the outer pixel values of FIG. 3E, the coefficients in the shaded portion of FIG. 3F form 0 or known values. Using run-length coding and variable-length coding, it is possible to further compress the selected L or smaller number of pixels and transform coefficients that become 0 in FIG. 3F.

Based on the flowchart in FIG. 3H , the selection and extrapolation processing of transform coefficients can be further explained.

An iterative POSC-based approach may be used to select the outer pixel values for a given object. In the first iteration, the outer pixel values can be set arbitrarily, which is known to be effective for iterating or reflecting the inner pixel values [S.F.Chang and D.G.Messerschmitt, "Transform Coding of an Arbitrarily-shaped Image Segment (arbitrarily shaped image Transform coding of segment) Proceedings of ACMMultimedia, August, 1993]. Once the external pixel value is selected, the rectangular block is transformed forward in order to obtain L or less than the number of transform coefficients. Since the size of each transform coefficient is the same as Corresponding to the energy associated with the coefficients, one method of coefficient selection is to select L (or a set number less than this number) largest transformation coefficients.

Once a coefficient is selected, other coefficients that are not selected are set to "0". Setting these coefficients to "0" will cause signal distortion in the spatial domain, so after the inverse transformation of the coefficients that are not set to "0" in the selected position, replace the pixel values inside and on the boundary of the object with the original values.

Only the outer pixel values are affected by forward transformation, inverse transformation and inner pixel replacement, and the coefficient values of unselected positions no longer become "0". Therefore, the previous process of "forward transformation → zeroing → inverse transformation → internal pixel replacement" is repeated until convergence. Known convergence can be guaranteed [e.g. H.H.ChenM.R.Civanlar and B.G.Haskell, "A Block Transform Coder for Arbitrarily Shaped Image Segments (block transform coder for arbitrary shape image segments)" Proceedings of IEEE International Conference on Image Processing, 1994, Vol.1, 85-89]. If it converges, the transform coefficients that have completed the conformal transformation can be further compressed as shown in FIG. 3F and FIG. 3G with the same example of run-level coding and variable-length coding. The transform coefficient blocks in FIG. 3F are scanned and run-level coded in zigzag scanning order. The [run, level] symbols can be further compressed with variable length encoding.

Figure 4 shows the escape and rule domains sorted by probability according to the [run, level] notation. Statistically speaking, the [run,level] symbol has a very low probability of having a large value for 'run' and/or 'level'. In the domain of low-probability symbol distribution, which is the escape domain, the symbols are expressed as fixed-length escape sequences, while regular Huffman codes are assigned to other domains, namely the regular domain. For example, in the case of 8×8 segments of data, the escape sequence consists of a 6-bit escape symbol, a 6-bit "stroke" from "0" to "63", an 8-bit "level" from "1" to "255" and 1 sign bit constitutes. Thus, the escape sequence has a fixed length of 21 bits in total.

In the existing variable-length coding system, the energy of the video signal is concentrated in the low-frequency region constituting the center of the AC (alternating current) component, so in the variable-length coding of video data, a zigzag is used for the N×N quantization coefficient Scan mode (see Figure 3A to Figure 3C). However, depending on the mode of the video signal, the energy of the video signal may be more widely distributed in the horizontal or vertical frequency components. Therefore, the existing zigzag scan mode is not the best scan mode for variable length coding of video data. Thus, for variable length encoding and variable length decoding, a scanning pattern inclined horizontally or vertically is desirable in order to have flexibility to vary with the distribution characteristics of the video data.

Fig. 5 shows a variable length coder according to an embodiment of the present invention. The coder shown in Fig. 5 is made up of following parts: respectively store the coefficient memory parts CM ₁ ～CM _N of the quantization coefficient of quantized segment data as shown in Fig. 3B and Fig. 3F quantized segment data; N scan address output parts SAG ₁ -SAG _N provided to the coefficient storage unit; N run-level encoders CD ₁ -CD _N that carry out process-level encoding for each coefficient of each coefficient storage unit according to the scanning mode; according to the available Variable-length coding mapping N variable-length coders VLC ₁ ～ VLC _N that perform variable-length coding on the [run, level] symbols output from each run-level coder; store each variable-length code of the variable-length coder N buffers BF ₁ to BF _N of data; N accumulators ACCM ₁ to ACCM _N respectively accumulating the length of the variable-length-coded data output from each variable-length coder; A minimum value selector for selecting the minimum value in the length accumulated by the device; a selection switch 54 for selecting and transmitting the buffer output of the variable length coding channel selected in the minimum value selector 52 .

First, quantized coefficients quantized into segments of a predetermined size are stored in N coefficient storage units CM ₁ to CM _N , respectively. The first, second, and Nth coefficient storage means receive the first, second, and Nth type scan addresses respectively outputted from the first, second, and Nth type scan address output means, respectively. The encoding channel of the first coefficient storage unit among the N coefficient storage units scanned according to N types of scan addresses will be described below.

Scan the quantized coefficients stored in the first coefficient storage unit CM ₁ in the set scanning direction according to the first type of scanning address, and encode the quantized coefficients in the first run-length layer coder CD ₁ into long [run-length, layer] symbols . The first variable length coder VLC ₁ performs variable length coding on the [run, level] symbols output from the first run level coder CD ₁ according to the set variable length code map, and outputs the variable length codes one by one The subsequent data D _VLC and the variable length coded data length L _VLC . The variable length coded data D _VLC output from the first variable length coder VLC ₁ is stored in the first buffer BF ₁ ; the length L _VLC of the variable length coded data is input Accumulation is carried out in the first accumulator ACCM ₁ of the long L _VLC coded by the VLC ₁ . The first accumulator _ACCM1 is composed of an adder _A1 and a first accumulation length storage unit _LM1 . In the adder A ₁ , the length L _VLC of the variable length coded data input from the first variable length coder VLC ₁ is added to the cumulative length fed back from the first cumulative length storage unit LM ₁ . The first cumulative length storage unit _LM1 stores the updated cumulative length output from the adder _A1 .

This coding pass consisting of a succession is applied to the quantized coefficients of the 2nd, 3rd and Nth coefficient storage means CM ₂ , CM ₃ , CM ₄ . However, in order to scan quantized coefficients of one segment stored in N coefficient storage sections one by one, another mode may be used. Fig. 7 shows a realization example of a plurality of other scanning modes. Figure 7A shows a scan pattern with a scan direction of 0 degrees, Figure 7B shows a scan pattern with a scan direction of 30 degrees, and Figure 7C shows a scan pattern with a scan direction of 45 degrees.

In the case of data segments corresponding to arbitrarily shaped objects and corresponding to image segments transformed by surrounding rectangles, the scanning pattern need not cover all frequency components.

In the variable length coding channel according to a plurality of scanning modes, the cumulative length data stored in the cumulative length memory parts of the N accumulators ACCM ₁ to ACCM _N are provided to the N minimum value selectors 52 one by one. At the input, a minimum value selector 52 determines the minimum value for the cumulative length. Respective output terminals of N buffers BF ₁ ˜BF _N storing data variable-length coded according to N types of scan patterns are respectively connected to N input terminals of the selection switch 54 . The minimum value selector 52 selects the minimum value from the cumulative length data input from the N cumulative length storage units LM ₁ to LM _N , respectively. The minimum value selector 52 outputs scanning mode data D _SCAN , which represents the scanning mode of the variable-length coding channel having the minimum cumulative length selected from the cumulative lengths, and the prescribed minimum value corresponding to the cumulative length is selected. The selection control signal SEL is supplied to the selection switch 54 . The selection switch 54 selects and outputs the variable-length-coded data D _VLC corresponding to the minimum value of _the accumulated length among the input data inputted one by one to N input terminals.

Whenever the minimum value is selected, that is, whenever the variable-length coding of each segment data is completed, the minimum value selector 52 generates a reset signal RST to reset the N buffers BF ₁ to BF _N and the N cumulative length storage units LM ₁ ~LM _N . Variable length coded data D _VLC and scan pattern data D _SCAN are output from the variable length coder as digital data which are either stored or transmitted to a receiver for decoding.

Fig. 6 shows an embodiment of a variable length decoder according to the present invention. Referring to FIG. 6, the variable-length coded data D _VLC input to the variable-length decoder 61 is converted into [run, level] symbols according to the variable-length decoding map. And, the scanning mode data D _SCAN that is sent from the decoder is input to the scanning mode selector 62, and this scanning mode selector 62 is used for storing and as shown in Figure 7 a plurality of scanning modes (the first to the Nth scanning) Corresponding to each scanning address. The scan mode selector 62 selects and outputs scan addresses ADDRs corresponding to the input scan mode data D _SCAN . The run-level decoder 63 converts the [run-level, level] symbols input from the variable-length decoder 61 into two-dimensional quantization coefficients based on the scan addresses ADDRs input from the scan mode selector 62 . Thereafter, the quantized coefficients are supplied to an inverse quantizer.

As mentioned above, according to the variable-length coding system of the present invention, the variable-length coding of each piece of data is performed according to the variable-length scanning mode or the scanning mode that minimizes the length of the variable-length coding data and the variable-length coding according to the scanning mode is transmitted. data, or store it on a digital recording medium for subsequent decoding. According to the variable length decoding system of the present invention, decoding is performed using the same scanning pattern as used in the variable length encoding process of stored or transmitted variable length encoded data. As a result, the system of variable length coding and variable length decoding can further compress transmission data.

In the present invention, each segment data is adapted to various sizes and shapes of image parts. That is, regardless of how digital data is divided into segment data, the present invention can be applied to encoding and decoding of segment data. This can be further understood based on FIGS. 3D to 3H . If the conformal transformation is done as described above, the transformation coefficients in the rectangular block are scanned in multiple scanning order, and then the specific scanning order is selected using the same method as illustrated in FIG. 5 . It is thus clear how the invention, described in connection with Fig. 5, can be extended to more general image portions of varying sizes and shapes.

The present invention has been illustrated and described above using a variable length encoder/decoder, but other types of encoder/decoders may also be used in the present invention. For example, a Huffman or arithmetic coder/decoder could be used instead. Furthermore, the advantage of the present invention is obtained with the most suitable scanning pattern irrespective of the type of encoder/decoder, therefore, the present invention can be used without a variable length encoder/decoder.

Furthermore, although the present invention has been illustrated and described in relation to two-dimensional data, the present invention is also applicable to encoding and decoding systems employing multi-dimensional data.

The extension to multi-dimensionality can be realized by co-scanning multiple segments. FIG. 8A shows a specific example of obtaining multiple segments from the same image. However, as shown as an example in FIG. 8B, other time segments may also be scanned simultaneously. As shown in FIG. 8C, the two cases of FIG. 8A and FIG. 8B can be mixed without any limitation. 8D and 8E show examples of scan patterns for multiple segments. The numbers shown in FIG. 8D and FIG. 8E represent the scanning order of the data segments in FIG. 3A or FIG. 3C.

The invention can often be applied to multidimensional situations without any restrictions.

Claims (30)

1. A method of encoding digital data in which an object within an image is segmented into segments, characterized in that the method comprises the following steps: obtaining segment data of the object, the segment data being a plurality of data segments corresponding to a plurality of objects; Obtain the transformation coefficient for the segment data of the above object; Scanning the data obtained in the above steps according to the selected scanning mode suitable for the segment data distribution characteristics of the object, and encoding the scanned data; outputting the scan mode data of the above coded data and the scan mode used to encode the data, wherein, Each of the segment data corresponds to an object selected within the image and indicates the size and shape of the image segment corresponding to the pixel value of the object. 2. The encoding method according to claim 1, characterized in that, in the encoding step, the scanning range according to the selected scanning mode only covers the distribution area of the segment of data. 3. The coding method as claimed in claim 1, characterized in that, in the step of obtaining the transform coefficients, the obtained data is a set coefficient, and corresponds to the data distribution characteristics of the object segment Consistent multiple two-dimensional data. 4. The encoding method according to claim 1, characterized in that, in the step of obtaining the transform coefficients, the obtained data is a plurality of two-dimensional data consistent with the distribution characteristics of the object segment data. Quantization coefficient. 5. The encoding method according to any one of claims 1 to 4, characterized in that, in the encoding step, variable length encoding is performed on the scanned data. 6. The encoding method according to any one of claims 1 to 4, characterized in that, in the encoding step, Huffman encoding is performed on the scanned data. 7. The encoding method according to any one of claims 1 to 4, characterized in that, in said encoding step, arithmetic encoding is performed on the scanned data. 8. The encoding method according to any one of claims 1 to 4, characterized in that, in the encoding step, the scanning direction of the selected scanning mode is the horizontal direction. 9. The encoding method according to any one of claims 1 or 4, characterized in that, in the encoding step, the scanning direction of the selected scanning mode is the vertical direction. 10. The encoding method according to any one of claims 1 or 4, characterized in that, in the encoding step, the selected scanning pattern has a scanning direction with an inclination of 30 degrees or 45 degrees. 11. The encoding method according to any one of claims 1 or 4, characterized in that, in the encoding step, the obtained data are respectively scanned and encoded according to a plurality of given scanning modes, and The scan mode corresponding to the minimum value among the accumulated length values of each coded data is the selected scan mode. 12. The encoding method according to claim 11, characterized in that, the given multiple scanning modes can include 0 degrees (horizontal), 30 degrees, 45 degrees (Z type) and 90 degrees (vertical) Scanning pattern in the direction of the obliquity. 13. A method of decoding encoded data, characterized in that the method comprises the steps of: receiving the above coded data for decoding; Arrange the above-mentioned decoded data according to the positions determined by the provided scan pattern data, wherein, The received encoded data corresponds to a plurality of data segments, The encoded data is segment-encoded data for objects within the image; Wherein, each of the segment data corresponds to an object selected in the image, and represents the size and shape of the image segment corresponding to the pixel value of the object. 14. The decoding method according to claim 13, characterized in that said scan mode data is provided together with the encoded data, and represents the scan information used by the encoded data. 15. The decoding method according to claim 13, characterized in that, in the arranging step, the data formed by arranging at predetermined positions is consistent with the distribution characteristics of each piece of data of the object before encoding. 16. The decoding method according to claim 13, wherein, in said decoding step, the received encoded data corresponds to a plurality of data segments of one picture. 17. The decoding method according to claim 13, characterized in that, in said decoding step, the received encoded data corresponds to a plurality of data segments of a plurality of images. 18. The decoding method according to claim 13, wherein the decoding step is to receive variable-length coded data and perform variable-length decoding on the coded data. 19. The decoding method according to claim 13, wherein the decoding step is to receive Huffman encoded data and perform Huffman decoding on the encoded data. 20. The decoding method according to claim 13, wherein the decoding step is to receive arithmetic coded data and perform arithmetic decoding on the coded data. 21. The decoding method according to any one of claims 13 to 20, characterized in that, in the arranging step, the scanning direction determined according to the scanning pattern data is the horizontal direction. 22. The decoding method according to any one of claims 13 to 20, characterized in that, in the arranging step, the scanning direction determined according to the scanning pattern data is the vertical direction. 23. The decoding method according to any one of claims 13 to 20, characterized in that, in the arranging step, the scanning direction determined according to the scanning pattern data is a scanning direction with an inclination angle of 30 degrees or 45 degrees . 24. The decoding method according to any one of claims 13 to 20, characterized in that, in the arranging step, a scanning address is generated corresponding to the scanning direction and scanning range determined according to the scanning pattern data, and in the The decoded data are arranged one by one at the positions specified by the scan address. 25. The decoding method according to claim 24, characterized in that, the generated scanning address is one of multiple scanning addresses selected according to the provided scanning mode data. 26. The decoding method according to claim 25, characterized in that, said multiple scanning addresses can include tilt angles of 0 degree (horizontal), 30 degrees, 45 degrees (Z-type) and 90 degrees (vertical) Direction scan address. 27. The decoding method according to any one of claims 13 to 20, characterized in that, in the step of arranging, the data obtained after arranging is a plurality of two-dimensional data. 28. The decoding method according to claim 27, wherein the two-dimensional data is two-dimensional data of a plurality of transformed coefficients. 29. The decoding method according to claim 27, wherein the two-dimensional data is a plurality of two-dimensional quantized coefficients. 30. The decoding method according to claim 29, further comprising the step of dequantizing the quantized coefficients.

Images