JP2013115528A - Video compression apparatus and video decompression apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a video compression technology, which adjusts image quality and a data amount more appropriately, and a video decompression technology corresponding thereto.SOLUTION: A video compression apparatus comprises a deterioration processing section which performs deterioration processing for deteriorating image quality of an input video image, an image quality measuring section which measures the image quality of the video image deteriorated by the deterioration processing section, and a reversible compression section which performs reversible compression on the video image of which the image quality has been deteriorated by the deterioration processing section. The deterioration processing section is configured to repeat the deterioration processing until the image quality measured by the image quality measuring section becomes predetermined image quality.

Description

The present invention relates to a video compression technique and a video restoration technique.

  A technique for reducing the amount of data by irreversibly compressing an input video signal is disclosed in Non-Patent Document 1, for example.

JP-A-7-203430

  However, Patent Document 1 does not disclose a specific configuration for a method of compressing while maintaining designated image quality. Also, there has been no disclosure of a method for keeping the code length below a certain value while maintaining the designated image quality.

  The present invention has been made in view of the above problems, and an object of the present invention is to provide a video compression technique in which image quality and data amount are adjusted more suitably, and a video restoration technique corresponding to the video compression technique.

  In order to solve the above problems, an embodiment of the present invention includes, for example, a degradation processing unit that performs a degradation process of image quality of an input video, and an image quality measurement that measures the image quality of a video that has been degraded by the degradation processing unit. And a reversible compression unit that reversibly compresses the video whose image quality has been degraded by the degradation processing unit, and the degradation processing unit performs the degradation process until the image quality measured by the image quality measurement unit reaches a predetermined image quality. What is necessary is just to make it the structure which repeats.

  ADVANTAGE OF THE INVENTION According to this invention, the video compression technique which adjusted the image quality and the data amount more suitably, and the video restoration technique corresponding to this can be provided.

It is a block diagram which shows the video compression transmission system which concerns on Example 1 of this invention. It is a block diagram which shows the lossless compression process part which concerns on Example 1 of this invention. It is a figure which shows the DPCM process which concerns on Example 1 of this invention. It is a figure which shows a response | compatibility with the numerical value which concerns on Example 1 of this invention, and a Golomb-Rice code. It is a block diagram which shows the reversible restoration process part which concerns on Example 1 of this invention. It is a block diagram which shows the video compression transmission system which concerns on Example 2 of this invention. It is a block diagram which shows the video compression transmission system which concerns on Example 3 of this invention.

  Hereinafter, preferred embodiments for carrying out the present invention will be described with reference to the drawings.

  First, Example 1 of the present invention will be described.

  FIG. 1 shows a video compression transmission system according to Embodiment 1 of the present invention. Hereinafter, the operation and details of each element will be described.

  The encoder 10 includes an image quality deterioration processing unit 101, a lossless compression processing unit 103, an image quality measurement processing unit 109, and a deterioration continuation determination processing unit 107, and generates a stream 104 in which the video signal 100 is compressed.

The image quality degradation processing unit 101 degrades the image quality of the video signal 100 and outputs the degraded video signal 102. The type (color, monochrome, visible light, infrared light) of the video signal 100 and the data structure are not limited. In this embodiment, as an example, the video signal 100 is a color image of three primary colors, arranged in a two-dimensional array of hundreds to thousands of pixels in the horizontal and vertical directions, and red, It is assumed that a green and blue component is assigned and a data structure that represents each color component with a depth of 8 bits. As an example of a specific image quality deterioration method in the image quality deterioration processing unit 101, typically, blurring processing using a Gaussian filter is performed. As will be described later, this deterioration process is repeatedly performed a plurality of times on the same video signal. As the number of times increases, it is desirable to select a degradation process that tends to shorten the code length described later. A filter corresponding to the image quality required for the video signal to be transmitted to the decoding side may be used as appropriate. For example, if compression is required while maintaining the sharpness of the image, an edge-preserving adaptive blur filter may be used. The image quality measurement processing unit 109 measures the image quality of the degraded video signal 102 based on the video signal 100 from the degraded video signal 102 and the video signal 100. For example, PSNR (Peak Signal to Noise Ratio) may be used as an index for image quality measurement. The PSNR may be obtained using Equation 1, for example. The PSNR value is output as image quality information 108.

(Equation 1)
MSE = Σ ((Io-In) × (Io-In)) / SIZE
PSNR = 20 × log10 (MAX / √MSE)
However, Io: Video signal
In: Degraded video signal
SIZE: Number of pixels in the image
MAX: Maximum depth of pixel value
Σ range: all pixels in the image

The deterioration continuation determination processing unit 107 uses the image quality information 108 as an input signal. The deterioration continuation determination processing unit 107 outputs a deterioration continuation necessity instruction signal 110 for instructing to stop the deterioration process if the image quality information 108 is equal to or less than a predetermined image quality threshold value. If the image quality information 108 is equal to or higher than the image quality threshold value, a deterioration continuation necessity instruction signal 110 for instructing continuation of the deterioration process is output. In other words, the image 102 satisfying the instructed image quality can be generated by continuing the deterioration process until a predetermined image quality is achieved. At this time, since the code length tends to be shorter as the image quality deterioration process is repeated, the stream 104 having a higher image quality than the predetermined value and a shorter code length is generated.

  The image quality degradation processing unit 101 repeats the degradation process several times for the same image while the degradation continuation necessity instruction signal 110 indicates the continuation of the degradation process. When the instruction to stop degradation processing is instructed by the degradation continuation necessity instruction signal 110, the image quality degradation processing unit 101 stops the repeated processing and outputs the degraded video signal to the lossless compression processing unit 103. The process proceeds to the processing by the lossless compression processing unit 103.

  The lossless compression processing unit 103 performs lossless compression encoding on the deteriorated video signal 102 and outputs it as stream data 104. FIG. 2 is a diagram illustrating an internal configuration of the lossless compression processing unit 103.

  A DPCM (Differential Pulse Code Modulation) processing unit 1030 calculates a difference value of the video signal 102 and outputs difference data 1031. The Golomb-Rice encoding processing unit 1032 performs code conversion based on the Golomb-Rice code on the difference data 1031 and outputs the result 104 as a stream 104.

  In this embodiment, DPCM and Golomb-Rice code are combined as lossless compression processing, but other lossless compression processing may be used. A reversible compression process that can be expected to have a higher compression rate in accordance with the characteristics of the input image and the image quality deterioration process may be employed.

  The DPCM processing unit 1030 calculates and outputs a difference between adjacent data. This is because, as a general characteristic of images and sounds, the correlation between adjacent data is high, and there is a high possibility that they are relatively close values, so that the difference value between them is expected to be small. FIG. 3 shows a specific example of DPCM processing. When 8-bit data of {125, 126, 129, 134, 130, 132, 115, 115} is input as the video signal 102, the difference data 110 is {-3, 1, 3, 5, -4. , 2, −17, 0}. Since the first data has no object for which the difference is taken, the constant 128 is used as the object of the difference here.

  Next, the Golomb-Rice encoding processing unit 1032 is a process of assigning a short code to a small value and a long code to a large value. As a result of the DPCM processing unit 1030, the difference data 1031 has a small value, and therefore it can be expected that the code length is shortened as a whole when Golomb-Rice coding is performed.

A specific method of Golomb-Rice encoding processing will be described. In Golomb-Rice encoding, an input numerical value n is encoded using a parameter k. First, p is calculated using Equation 2, code 0 is added by the number of p, and code 1 is then added.

(Equation 2)
p = FLOOR ((| n | +1) / k) -1
However, FLOOR () rounds down after the decimal point.

Next, a sign 0 is added when the numerical value n is positive, and a sign 1 is added when it is negative.

Furthermore, when q is calculated using Equation 3 and added as a code, encoding is completed.

(Equation 3)
q = (| n | +1) mod k−1

As an exception, when the numerical value is 0, the sign is 1.

  FIG. 4 is a table showing an example of Golomb-Rice coding, and the parameter k is 2. From this table, for example, it is understood that the code corresponding to the numerical value −3 is “00110”.

  When {-3, 1, 3, 5, -4, 2, -17, 0}, which is an example of the difference data 110 described above, is Golomb-Rice encoded, {00110,0100,00100,000100,00111,0101, 000000001110,0}, and the code lengths thereof are {5, 4, 5, 6, 5, 4, 12, 1}, respectively. Therefore, the total code length is 42 bits. Since the data length of the input video signal is 8 pixels at 8 bit depth, it is 64 bits, which means that 22 bits can be compressed.

Note that the code length of the Golomb-Rice code can be easily obtained using Equation 4. Here, k is a parameter of Golomb-Rice code, and n is a numerical value to be encoded.

(Equation 4)
if k = 0, 1
others, FLOOR ((| n | +1) / k) + 1 + k
However, FLOOR () rounds down after the decimal point.

Stream data 104 is generated by the Golomb-Rice coding described above. In the example of FIG. 1, the stream data 104 is transmitted from the encoder 10 to the decoder 11.

  The decoder 11 includes a lossless restoration processing unit 105 and restores the video signal 106 from the stream 104.

  The lossless restoration processing unit 105 is a process of decoding the losslessly compressed code (reversing the lossless compression) with respect to the input stream 104 to restore the video signal 106. Here, the inverse DPCM process is performed after the Golomb-Rice decoding process so as to correspond to the compression process of the lossless compression processing unit 103.

  FIG. 5 is a diagram illustrating an internal configuration of the reversible restoration processing unit 105. First, the Golomb-Rice decoding processing unit 1050 performs Golomb-Rice decoding of the stream 104 and outputs it as difference data 1051. Next, the inverse DPCM processing unit 1052 restores the video signal 106 by calculating the cumulative addition of the difference data.

  The Golomb-Rice decoding processing unit 1050 performs Golomb-Rice decoding on the input stream 104. Specifically, the Golomb-Rice code can be decoded according to the following procedure. As a premise, it is assumed that the parameter k at the time of encoding is known.

First, count how many 0s continue until 1 comes to the code, and let this number be m. When m is 0, the decoded numerical value is 0. The next sign indicates the sign of the numerical value. Here, -1 is substituted for f if 1 and 1 is substituted if 0. Subsequently, the code for k-1 bits is read, and when this is set to a, the numerical value n is obtained using Equation 5.

(Equation 5)
n = f × (m × k + a−1)

For example, when the code is 00110, m is 2 because two 0s continue until 1 comes first. Since the next sign is 1, the numerical value is negative, and −1 is substituted for f. Finally, since 1 (m−1) bits are read as 0, a is 0. When these numerical values are substituted into Equation 5, n is −3, and the decoded numerical value is −3.

  The differential data 1051 is restored by the Golomb-Rice decoding process described above.

  The inverse DPCM processing unit 1052 calculates the cumulative addition of the difference data 1051 restored by the Golomb-Rice decoding process, and restores the video signal 106.

  For example, an example in which the video signal 106 is restored when the difference data 1051 is {-3, 1, 3, 5, -4, 2, -17, 0}.

  First, when the constant 128 used in the DPCM processing unit 1030 and the difference data-3 are added, the first pixel value is obtained as 125. Next, 125 which is the addition result and 1 of the difference data are added, and the second pixel value becomes 126. If the pixel value is obtained by repeating the addition of the difference values in this way, the video signal 106 can be restored. In this example, {125, 126, 129, 134, 130, 132, 115, 115} is restored as the video signal 106. This is exactly the same as the video signal 102 input to the lossless compression of the encoder 10, and the lossless compression processing unit 103 and the lossless restoration processing unit 105 can perform the lossless compression / decompression of the video signal 102.

  With the above configuration, it is possible to compress an input image into an image having a predetermined code length and a short code length. According to the first embodiment of the present invention described above, video compression in which image quality and data amount are adjusted more appropriately, and video restoration corresponding to this can be realized.

  Next, a second embodiment of the present invention will be described.

FIG. 6 is a processing block diagram of the video compression transmission system according to the second embodiment of the present invention. Hereinafter, the operation and details of each element will be described. Note that the operations of the image quality deterioration processing unit 101, the lossless compression processing unit 103, the image quality measurement processing unit 109, the decoder 11, and the lossless restoration processing unit 105 of the encoder 10 are the same as those described in the first embodiment, and therefore the description will be made. Omitted.
In addition to the image quality degradation processing unit 101, the lossless compression processing unit 103, the lossless restoration processing unit 105, the image quality measurement processing unit 109, and the degradation continuation determination processing unit 107 described in the first embodiment, the encoder 10 performs, for example, a code length measurement process. The unit 112 is provided. The video signal 100 is input to the encoder 10 to generate a compressed stream 104.

  The code length measurement processing unit 112 calculates the code length when the degraded input image 102 is compressed by the lossless compression processing unit 103 based on the difference data 1031 calculated by the lossless compression processing unit 103, and outputs the code length 111. It is processing to do.

  Specifically, the total code length of the difference data 1031 is calculated using Equation 4 to obtain the code length 111.

  As described above, since the Golomb-Rice code can obtain the code length before actually encoding, when the Golomb-Rice code is adopted for the lossless compression processing unit 103, the input of the code length measurement processing unit 112 May be difference data 1031. However, when another lossless compression method is adopted for the lossless compression processing unit 103, the code length may not be measured unless the encoding is actually performed. In this case, the stream data 104 is input to the code length measurement processing unit 112, the data length of the encoded data is measured, and the code length 111 is obtained.

  The image quality information 108 and the code length 111 are input to the deterioration continuation determination processing unit 107. The deterioration continuation determination processing unit 107 outputs a deterioration continuation necessity instruction signal 110 for instructing to stop the deterioration process when the code length 111 is equal to or less than a code length threshold specified in advance. If the code length 111 is equal to or greater than the code length threshold, a deterioration continuation necessity instruction signal 110 that instructs to continue the deterioration process is output. At this time, when the image quality information 108 falls below a predetermined image quality threshold, it may be determined that compression is impossible with the predetermined image quality threshold and code length threshold, and the processing may be stopped. Therefore, a stream including compressed video information having a picture length equal to or higher than a predetermined threshold and a code length shorter than a predetermined code length is output to a transmission destination device.

  With the above configuration, it is possible to generate a stream having a code length equal to or less than a predetermined threshold while satisfying a predetermined image quality.

  According to the second embodiment of the present invention described above, a stream satisfying a predetermined image quality and a predetermined code length can be generated more suitably. Thereby, it is possible to realize video compression in which the image quality and the data amount are adjusted more suitably, and video restoration corresponding to the video compression.

  Next, Embodiment 3 of the present invention will be described.

  FIG. 7 is a processing block diagram of the video compression transmission system according to the third embodiment of the present invention. Hereinafter, the operation and details of each element will be described. Note that the operations of the image quality degradation processing unit 101, the image quality measurement processing unit 109, and the degradation continuation determination processing unit 107 of the encoder 10 are the same as those described in the first embodiment, and thus are omitted. The decoder 11 includes an image quality restoration processing unit 115 in addition to the lossless restoration processing unit 105 described in the first embodiment, and inputs the compressed stream 104 and outputs a video signal 116.

  In addition to performing the deterioration process described in the first embodiment, the image quality deterioration processing unit 101 of the encoder 10 outputs the number of repetitions of the deterioration process as the image quality deterioration processing information 113. In addition, the image quality deterioration processing information 113 may store information for identifying the image quality deterioration processing performed by the image quality deterioration processing unit 101, such as filter characteristics used for the image quality deterioration processing. The lossless compression processing unit 103 includes the image quality degradation processing information 113 in the stream 104 in addition to performing the compression processing described in the first embodiment. For example, a stream including deterioration information in the header portion and an encoded and compressed video signal in the data portion is used.

  In addition to performing the restoration process described in the first embodiment, the reversible restoration processing unit 105 extracts image quality degradation processing information from the stream 104 and outputs the extracted image quality degradation processing information 114.

  The image quality restoration processing unit 115 performs restoration processing of the degraded video signal 106 based on the image quality degradation processing information 114.

  First, it is identified from the image quality degradation processing information 114 what frequency distribution filter the image quality degradation is performed. When the image quality degradation processing unit 101 of the encoder 10 repeats the Gaussian filter a plurality of times, the image quality degradation processing information 114 includes identification information indicating the number of times the Gaussian filter is repeated or the filter characteristics of other Gaussian filters repeated a plurality of times. Is stored and transmitted. Therefore, based on the image quality degradation processing information 114, the filter characteristics of the degradation processing performed by the image quality degradation processing unit 101 of the encoder 10 can be identified. The reciprocal of the identified filter characteristic is multiplied by the Fourier transform of the video signal 106, and then the inverse Fourier transform is performed to obtain a restored video signal 116. That is, the image quality restoration processing unit 115 identifies the image quality degradation processing performed by the image quality degradation processing unit 101 of the encoder 10 based on the image quality degradation processing information 114, and performs the image quality restoration processing corresponding thereto. The restored video signal becomes a high-quality video signal similar to the input image 100 of the encoder 10 than the degraded video signal 106.

  According to the third embodiment of the present invention described above, video compression in which the image quality and the data amount are adjusted more appropriately, and video restoration corresponding to this can be realized, and a restored image closer to the image quality of the original image is obtained more suitably. be able to.

10 Encoder 11 Decoder 101 Image quality degradation processing unit 103 Lossless compression processing unit 105 Lossless restoration processing unit 107 Degradation continuation determination processing unit 109 Image quality measurement processing unit 112 Code length measurement processing unit 115 Image quality restoration processing unit 1010 DPCM processing unit 1032 Golomb-Rice code Processing unit 1050 Golomb-Rice decoding processing unit 1052 Inverse DPCM processing unit

Claims (8)

A degradation processing unit that performs degradation processing of the image quality of the input video;
An image quality measuring unit for measuring the image quality of the video whose image quality is degraded by the degradation processing unit;
A reversible compression unit that reversibly compresses the video whose image quality has been degraded by the degradation processing unit,
The video compression apparatus, wherein the deterioration processing unit repeats the deterioration processing until the image quality measured by the image quality measurement unit reaches a predetermined image quality. The video compression apparatus according to claim 1,
The video compression apparatus outputs a stream including a video reversibly compressed by the lossless compression unit and information for identifying image quality degradation processing performed by the degradation processing unit. The video compression apparatus according to claim 1,
The video compression apparatus, wherein the image quality measurement unit measures the image quality by comparing an image whose image quality is deteriorated with an input image by the deterioration processing unit. The video compression apparatus according to claim 1,
The video compression apparatus, wherein the image quality measurement unit calculates a PSNR of the video whose image quality has been deteriorated by the deterioration processing unit to obtain a measured image quality. A degradation processing unit that performs degradation processing of the image quality of the input video;
A reversible compression unit that reversibly compresses the video whose image quality has been degraded by the degradation processing unit;
A code length measurement unit that measures the code length of the image information output by the lossless compression unit,
The video compression apparatus, wherein the deterioration processing unit repeats the deterioration processing until the code length measured by the code length measurement unit becomes smaller than a predetermined code length. The video compression apparatus according to claim 5,
And an image quality measuring unit for measuring the image quality of the video whose image quality has been degraded by the degradation processing unit,
The lossless compression unit outputs a stream including compressed video information in which an image quality measured by the image quality measurement unit is equal to or higher than a predetermined image quality and a code length measured by the code length measurement unit is smaller than a predetermined code length. A video compression device. An input unit for inputting compressed video that has undergone image quality degradation processing and lossless compression processing;
A video decompression apparatus comprising: a compression / decompression processing unit that performs a reverse process of the lossless compression performed on the compressed video input to the input unit. The video restoration device according to claim 7,
The input unit receives identification information for identifying the image quality degradation process together with the compressed video,
And an image quality restoration processing unit for restoring the image quality of the video that has been subjected to the reverse processing of the lossless compression in the compression restoration processing unit,
The video image restoration apparatus, wherein the image quality restoration processing unit performs an image quality restoration process corresponding to the image quality deterioration process identified by the identification information.

Images