JPH11205800A - Picture improving device for encoding picture - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a picture improving device preventing the deterioration of the predicting efficiency of and I-picture and after it and the deterioration of picture. SOLUTION: Quantization step information with respect to an encoding mode (d) is outputted to a comparing and judging part 6 from a picture encoding part 5 corresponding to the I-picture, P-picture and B-picture of an inputted original picture signal (a). Quantization step information corresponding to the I-picture of much high-frequency information is more than a prescribed threshold value, the part 6 outputs a selection signal (f) to a selector 3 to give band limiting processing deleting high-frequency information of the signal (a) to make a band limiting picture (b) an encoding picture (c) to send to a picture encoding part 5 and when the mode (d) corresponds to the P-picture and B- picture, the selector 3 selects the signal (a) to execute efficient encoding processing according to the mode (d) and outputs a picture stream (g).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention provides a low-pass filtering process for reducing a high-frequency region of an input original image signal when a quantization parameter of an I-picture becomes equal to or more than a threshold value. The present invention relates to an image quality improvement apparatus for image coding that does not include area information and avoids a reduction in prediction efficiency after an I picture.

[0002]

2. Description of the Related Art Japanese Patent Application Laid-Open No. 61-216 discloses a technique for encoding an image signal used for transmitting the image signal.
No. 581 discloses that an input image signal is input to an encoding device through a low-pass filter during sample encoding of an input image. Also, JP-A-61-56
No. 521, a component removed by a low-pass filter for sampling is gradually reduced before sampling is turned on.
A technique has been disclosed in which the gain of the wideband component is gradually reduced from 1 to 0 as the amount of accumulation in the buffer memory increases, and the highband component is reduced to 0 or almost 0 in a coding mode before sampling is performed. I have.

[0003] Furthermore, an encoding circuit that performs encoding by inputting an original image signal containing high-frequency information, which is difficult to encode, without using a low-pass filter is also conceivable. When the original image signal is input to the encoding circuit, both high-frequency information and low-frequency information are to be encoded. For this reason, coding information is allocated to the low band and the high band information, and large quantization noise is superimposed on the low band information, and the high band information has a low degree of redundancy. In that case, the effect is not so much expected. Conversely, the quantization noise superimposed on the low-frequency information increases the prediction error signal, resulting in a reduction in prediction efficiency.

[0004] By the way, when performing high-efficiency encoding of an image signal, an I-picture for performing only intra-coding, a P-picture and a B-picture for performing inter-coding are periodically repeated. In general, an I picture has a large amount of generated information. For this reason, when an image whose encoding is strict is input, quantization is coarsened to suppress the amount of generated information of the I picture, and wide area information is reduced. Also, since the original picture signals of the P picture and the B picture following the I picture contain high frequency information, the prediction error signal component using the I picture as a reference picture contains the high frequency information as it is. Will be.

[0005]

For this reason, when the amount of encoded information of an image signal is increased, if the amount of encoded information is assigned to high-frequency information, the quantization step becomes coarser as a whole, and High-frequency information is reduced. As a result, the high-frequency information included in the input original image remains in the prediction error signal as it is, and by allocating the amount of coded information to the high-frequency information, the quantization step becomes increasingly coarse, resulting in a vicious circle. . As a result, there is a problem that image quality is deteriorated.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned conventional problems, and it is possible to avoid a reduction in the encoding efficiency of an image signal and to visually recognize an image code that prevents image quality deterioration. It is an object of the present invention to provide an image quality improving device.

[0007]

To achieve the above object, an image quality improving apparatus for image encoding according to the present invention performs a low-pass filtering process on an input original image signal and outputs a low-pass image signal. A filter, an encoding mode determining unit for determining an image encoding mode, and encoding of one of the input original image signal and the low band image signal according to the encoding mode determined by the encoding mode determining unit. Performing a high-efficiency encoding process on the image and outputting an image stream, an image encoding unit that outputs quantization step information, and the image encoding unit according to the quantization step information output from the image encoding unit And a coded image selection unit for selecting one of the input original image signal and the low-frequency image signal and outputting the coded image to the image coding unit.

According to the present invention, the coded image selecting means sets the input original image signal in accordance with the quantization step information output from the image coding section in accordance with the coding mode determined by the coding mode determining means. Alternatively, one of the low-pass image signals subjected to the low-pass filtering processing is selected, and the encoded image is transmitted to the image encoding unit. The coded image section performs high-efficiency coding on the coded image according to the coding mode determined by the coding mode determining means, and outputs an image stream.

[0009]

Next, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a first embodiment of the present invention. In the embodiment shown in FIG. 1, when the effect of avoiding the reduction of the coding efficiency does not satisfy the image quality of the I picture, the prediction efficiency of the P picture used as the prediction reference image is reduced in order to reduce the prediction efficiency. , Low-pass filtering is applied to the original image signal to reduce high-frequency information, and priority is given to encoding of low-frequency information, which is important for encoding prediction, so that image quality above a certain level is maintained. is there. In FIG. 1, an input original image signal such as a television signal (hereinafter referred to as an original image signal a) is input to an input terminal 1, and the original image signal a is supplied to a low-pass filter 2 and a selector 3. To be entered.

The low-pass filter 3 is for removing high-frequency components of the original image signal a. That is, an I picture that performs only intra coding of the original image signal a, and an I picture among B pictures and P pictures that perform inter coding generally have a large amount of generated information. In order to prevent the image from becoming coarse, the high-frequency component of the original image signal a is removed. Either the band-limited image signal b from which the high-frequency component has been removed by the low-pass filter 3 or the original image signal a introduced to the input terminal 1 is selected by the selector 3 and the encoded image c
Is transmitted to the image encoding unit 5.
The coding mode d from the I, P, B determining unit 4 as coding mode determining means is input to the image coding unit 5. When performing high-efficiency encoding of the original image signal a, the I, B, and P determining unit 4 performs an I-picture that performs only intra-encoding, a P-picture that performs inter-encoding,
When a picture is repeated periodically, the original image signal a
Is determined to be an I-picture, P-picture, or B-picture encoding mode, and the encoding mode d is output to the image encoding unit 5.

The image encoding unit 5 outputs the quantization step information e to the comparison / determination unit 6 in accordance with the encoding mode d, and outputs the quantization step information e to the encoded image c sent from the selector 3. The image stream g is sent to the output terminal 7 after performing the high efficiency coding process. Upon receiving the quantization step information e, when the encoding mode d is an I picture, the comparison determination unit 6 sets the quantization step information e to a predetermined upper threshold Th.
A comparison is made as to whether it is greater than or equal to 1 or less than or equal to a lower limit threshold Th2, and a selection signal f is output to the selector 3 according to the result of the comparison.

In this case, when the quantization step information e is equal to or more than the upper threshold value Th1, a selection signal f for turning on the low-pass filter 2 is output to the selector 3, and conversely, When the quantization step information e is equal to the lower limit threshold value Th2
In the following cases, a selection signal f that turns off the low-pass filter 2 is output to the selector 3. The selector 3 and the comparison / determination unit 6 constitute an encoded image selection unit 8. When the selector 3 turns on the low-pass filter 2 in response to the selection signal f from the comparison / determination unit 6, the output signal of the low-pass filter 2, that is, the band-limited image from which the high-frequency component of the original image signal a has been removed. b is selected and transmitted to the image encoding unit 5 as an encoded image c. Conversely, when the selector 3 turns off the low-pass filter 2 based on the selection signal f from the comparison / determination unit 6, the selector 3 selects the original image signal a input from the input terminal 1 and c is sent to the image encoding unit 5.

Next, the operation of the first embodiment configured as described above will be described with reference to the explanatory diagram of FIG. An input terminal 1 receives an original image signal a such as a television signal. The original image signal a is input to the selector 3 and also to the low-pass filter 2. In order to perform the high-efficiency encoding of the original image signal a, FIG. 2A shows an I picture for which only intra-coding, a P-picture and a B-picture for which inter-encoding are performed, of the original image signal a. Repeat periodically as shown. Of the I picture, P picture, and B picture of the original picture signal a, the I picture generally has a large amount of generated information.

As shown in FIG. 2A, the I, P, B deciding section 4 decides the coding mode d corresponding to the I picture, P picture, and B picture. This encoding mode d
The encoding mode d is sent from the I, P, B determining unit 4 to the image encoding unit 5. The image encoding unit 5 outputs the quantization step information e corresponding to the encoding mode d to the comparison determination unit 6. This quantization step information e is obtained by coding mode d
, And the encoding mode d corresponds to an I picture, a P picture, and a B picture, as is clear from the above description.

As shown in FIGS. 2A and 2B, corresponding to the I picture, the P picture, and the B picture,
The quantization step information corresponding to the picture and the B picture is between the upper threshold value Th1 and the lower threshold value Th2 set by the comparison / determination unit 6, and the generated information amount is large in the I picture, As shown in FIG. 2B, the upper limit threshold value Th1 is exceeded, or the lower limit threshold value T
h2 or less. Such an encoding mode d
Is determined by the comparison / determination unit 6 between the upper threshold value Th1 and the lower threshold value Th2, or exceeds the upper threshold value Th1, or exceeds the lower threshold value Th1. Is compared with the threshold value Th2, and the selection signal f is output from the comparison / determination unit 6 to the selector 3
Output to

If it is determined that the encoding mode d corresponds to the I-picture and that the quantization step information e exceeds the upper limit threshold value Th1 as a result of the determination by the comparison determination section 6, the selection signal f outputs to the selector 3 as a selection signal f for turning on the low-pass filter 2. As a result, the selector 3 turns on the low-pass filter 2 as shown in FIG. Therefore, while the low-pass filter 2 is on, the low-pass filter 2 outputs to the selector 3 as a band-limited image b that has been subjected to band-limiting processing for reducing high-frequency information of the original image signal a. The band-limited image b is input to the selector 3 and is sent from the selector 3 to the image encoding unit 5 as an encoded image c.

The image encoding unit 5 performs a high-efficiency encoding process on the encoded image c in accordance with the encoding mode d, and outputs the result to the output terminal 7 as an image stream g. In this case, the encoding mode d for the encoded image c in the image encoding unit 5 corresponds to the I picture, and therefore, high-frequency information that is difficult to encode in the original image signal a is deleted, and the low-frequency information is deleted. Only the information is encoded, and high-efficiency encoding becomes possible.

As shown in FIG. 2B, the quantization step in the coding mode d corresponding to the P picture and the B picture between the upper threshold value Th1 and the lower threshold value Th2. Also in the case of the information e, the selection signal f output from the comparison / determination unit 6 is sent to the selector 3, and the selector 3 turns on the low-pass filter 2 in the same manner as described above.
However, in this case, the original image transmission signal a does not include the high band information, and the low band filter 2 does not need to perform the band reduction process for the high band information.
In addition to the output signal that has passed through, the original image signal a is selected by the selector 3 and sent to the image encoding unit 5 as an encoded image c.

In this case, the image encoding unit 5 performs a high-efficiency encoding process on the encoded image c according to the encoding mode d, and outputs the result to the output terminal 7 as an image stream g. In this case, the encoding mode d for the encoded image c in the image encoding unit 5 is P picture other than I picture, B picture
It corresponds to a picture, and therefore, only low-band information that is easy to encode in the original image signal a is encoded, enabling high-efficiency encoding and encoding without reducing encoding prediction efficiency. Becomes

Next, although the encoding mode d corresponds to the I picture, as shown in FIG.
Is equal to the lower limit threshold value Th.
If the value is equal to or less than 2, the selection signal f output from the comparison determination unit 6 to the selector 3 is output to the selector 3 as a selection signal for turning off the low-pass filter 2. Therefore, the selector 3 turns off the low-pass filter 2. When the low-pass filter 2 is turned off, the selector 3 outputs the original image signal a introduced to the input terminal 1.
Is selected and sent to the image encoding unit 6 as the encoded image c.

In this case as well, the image encoding section 6 performs high-efficiency encoding processing in accordance with the encoding mode d determined by the I, P, B determining section 4 and outputs it to the output terminal 7 as an image stream g. When the low-pass filter 2 is off, the original image signal a does not include high-frequency information, and there is no need to encode high-frequency information. Only the low-frequency original image signal important for encoding prediction is encoded. Encoding mode d is a P picture,
Encoding can be performed on a B-picture image without lowering the prediction efficiency. Therefore, in any case of the encoding process, it is possible to use a high-efficiency code without deteriorating the image quality of the original image signal.

[0022]

As described above, according to the present invention, when the quantization parameter of an I picture exceeds a predetermined threshold, the high-frequency information of the original image signal is reduced by the low-frequency filter. Since the encoding process is performed on the encoded image that has been subjected to the low-pass filtering process, it is possible to give priority to the encoding of the low-frequency information of the original image signal that is important for the encoding prediction so that the image quality is equal to or higher than a certain level. This has a visually recognizable effect that the deterioration of the image quality can be prevented.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a first embodiment of an image quality improving apparatus for image encoding according to the present invention.

FIG. 2 is an explanatory diagram for explaining an operation of the image quality improving apparatus for image encoding in FIG. 1;

[Explanation of symbols]

1 ... input terminal, 2 ... low-pass filter, 3 ... selector, 4 ... I, P, B determination unit, 5 ... image coding unit, 6 ...
... comparison / determination section, 7 ... output terminal, 8 ... encoded image selection means.

Claims (6)

[Claims] A low-pass filter that performs low-pass filtering on an input original image signal to output a low-pass image signal; a coding mode determining unit that determines an image coding mode; and the coding mode determining unit. According to the determined encoding mode, a high-efficiency encoding process is performed on one of the encoded image of the input original image signal and the low-band image signal to output an image stream, and quantization step information is output. An image encoding unit to be output, and selecting one of the input original image signal and the low-frequency image signal in accordance with the quantization step information output from the image encoding unit, to encode the encoded image. An image quality improving apparatus for image encoding, comprising: an encoded image selecting unit that outputs to an image encoding unit. 2. The coding mode determining means according to claim 1, wherein said coding mode determining means performs only intra coding at the time of high efficiency coding of the input original image signal.
2. The image quality improving apparatus according to claim 1, wherein the coding mode is determined by determining which of a picture and a B picture or a P picture to be inter-coded to perform predictive coding. 3. The apparatus according to claim 1, wherein the low-pass filter deletes high-frequency information of the input image and outputs a low-frequency band-limited image. 4. A comparing and judging unit that compares the quantization step information output from the image encoding unit with a predetermined threshold value and outputs a selection signal,
A selector that selects one of the input original image signal and the low-frequency image signal according to the selection signal output from the comparison determination unit and outputs the encoded image to the image encoding unit. 2. The image quality improving apparatus for image encoding according to claim 1, wherein: 5. The comparison / determination unit switches the low-pass filter to an ON side when the quantization step information becomes equal to or more than an upper threshold in the coding mode in the case of an I picture of the input original image signal. 5. The image code according to claim 4, further comprising: outputting a power selection signal; and outputting a selection signal to switch the low-pass filter to an off side when the quantization step information becomes equal to or less than the lower threshold. Image quality improvement device. 6. The selector, when the quantization step information is equal to or more than the upper limit threshold, turns on the low-pass filter by the selection signal to generate a low-band band-limited image of the input original image signal. 5. The image quality improving apparatus according to claim 4, wherein the image quality is improved.