JP6460691B2 - Video compression apparatus, video compression method, and video compression program - Google Patents

Description

  The present invention relates to a video compression apparatus, a video compression method, and a video compression program for compressing a data amount of a high-definition video signal.

In recent years, research and development of super high-definition (resolution 7680 × 4320) having 16 times the number of pixels of high-definition as a high-definition video system exceeding high-definition (resolution 1920 × 1080) has been promoted.
Super Hi-Vision has an extremely large amount of data compared to conventional Hi-Vision, so large-scale signal processing devices and high-speed interfaces are required for content production. Therefore, in order to realize more compact and agile content production, a dual green super hi-vision signal is used (see Non-Patent Document 1).

  As shown in FIG. 14A, a dual green type super high-definition signal (hereinafter referred to as a DG signal) is composed of a Bayer array frame image known as a color filter array of a single-plate image sensor. In the Bayer array, four types of pixels of red (R), green (G1 and G2), and blue (B) are arranged at different positions. Compared to the super high-definition signal (b) with full resolution (about 100 million pixels of full color), the DG signal has 1/3 the number of pixels (about 33 million pixels). It is advantageous for saving. The DG signal is used as an image sensor or an interface signal between devices in many Super Hi-Vision content productions.

By the way, as an encoding method for compressing the data amount of the video signal, for example, the image data is divided into a plurality of low resolution blocks, and two-dimensional discrete cosine transform (DCT) is performed on a block basis, and the obtained DCT There is a method of quantizing coefficients and performing entropy coding. As a representative example of such a technique, FIG. 15 shows a JPEG processing procedure.
When such an existing encoding method is applied to a DG signal, the frame image of the Bayer array is converted into a full color image (demosaicing), or the frame image of the Bayer array is converted into R, G1 as shown in FIG. , G2, and B are divided into pixels (sub-sampling), reconstructed into four 4K images (resolution 3840 × 2160), and then encoded (compressed) (Patent Document 1 and Non-Patent Document). 2).

JP 2006-019804 A

Tomoyuki Yamashita, Koji Mitani, Masayuki Sugawara, Hiroshi Shimamoto, Fumio Okano: “4000 Scanning Lines 4-Plate Ultra High Definition Video Camera”, Journal of the Institute of Image Information and Television Engineers, Vol. 58, no. 3, pp. 383-391 (March 2004) E. Miyashita and T.M. Kajiyama: “Compact camera recorder for Super Hi-Vision”, IEEE International Conference on Consumer Electronics 2014, pp. 165-166 (January 2014)

  However, in the method of performing demosaicing, complicated calculation is required to accurately estimate the insufficient pixels, and the load on hardware increases. Furthermore, since the number of pixels is tripled, the amount of data before compression increases.

  Also, in the method of pixel division, two G signals G1 and G2 having a spatially continuous correlation are spatially divided and encoded separately, so that adjacent pixels are separated at the boundary portion of the divided image. There are problems such as loss of continuity with pixels, reduction in compression efficiency, and compression distortion. In addition, when this method is used in a video recording apparatus, a single frame image is divided and recorded into a plurality of recording data, and therefore recording on a plurality of recording media or a plurality of recording areas is required. Management, browsing of recorded data, and backup work become complicated. Furthermore, when reproducing a DG signal recorded by such a video recording apparatus, a timing synchronization circuit for reading the divided recording data in parallel is required, which causes a problem that the apparatus configuration becomes complicated. there were.

  An object of the present invention is to provide a video compression apparatus, a video compression method, and a video compression program that can efficiently encode a high-definition video signal.

  The video compression apparatus according to the present invention includes a pixel dividing unit that divides a pixel for each color component and a first divided by the pixel dividing unit for a dual-green single frame image having a Bayer array sampling structure. The color component image is divided by a pixel interpolation unit that interpolates by synthesizing an image obtained by converting each pixel of the image according to a geometric rule, the first color component after interpolation, and the pixel division unit. An encoding unit that performs encoding by a predetermined compression method for each processing unit in which the second and third color components are packaged.

According to this configuration, the video compression apparatus generates a full-resolution G signal by performing an interpolation process having a geometric feature on the first color component (G signal) of the Bayer array frame image. To do. Further, the video compression apparatus packages the signal obtained by using such interpolation processing into a processing unit defined by a predetermined compression method and performs compression coding.
Therefore, since the video compression apparatus can compress a high-definition video signal using an existing encoding method, the cost of encoding hardware can be suppressed. In addition, since the video compression apparatus can compress the frame image while maintaining the continuity of the image without spatially dividing the frame image into a plurality of images, the compression distortion can be suppressed. Furthermore, the video compression apparatus employs a pixel interpolation method having a geometric feature, so that a two-dimensional histogram after spatial frequency conversion is converted into a specific component (u-axis direction, v-axis direction or diagonal direction). Therefore, the compression efficiency by entropy encoding is improved, and a high-definition video signal can be encoded efficiently.

  The video compression apparatus may include a blocking unit that divides the single frame image into blocks suitable for the processing unit, and the pixel interpolation unit may perform interpolation processing for each block.

  According to this configuration, since the video compression apparatus divides into processing block units of the encoding scheme and performs the interpolation processing for each block, the encoding efficiency can be improved by the interpolation processing adapted to the specified encoding scheme.

  The pixel interpolation unit may interpolate the image of the first color component by synthesizing it with an image obtained by inverting each pixel of the image in the horizontal direction.

  According to this configuration, the video compression apparatus can easily create the geometric feature after the interpolation processing by the image conversion that is reversed in the horizontal direction, and can improve the encoding efficiency.

  The pixel interpolation unit may interpolate the first color component image by synthesizing it with an image obtained by inverting each pixel of the image in the vertical direction.

  According to this configuration, the video compression apparatus can easily create the geometric feature after the interpolation processing by the image conversion inverted in the vertical direction, and can improve the encoding efficiency.

  The pixel interpolation unit may interpolate the first color component image by synthesizing the first color component image with an image obtained by rotating each pixel of the image about a central point by 90 degrees.

  According to this configuration, the video compression apparatus can easily create the geometric feature after the interpolation processing by image conversion rotated by 90 degrees, and improve the encoding efficiency.

  The video reproduction device according to the present invention is the video compression device according to any one of claims 1 to 5, wherein the decoding unit decodes the video data encoded by the predetermined compression method, and the decoding unit. Of the decoded image data, the color component separation unit that separates the first, second, and third color components, and the pixel of the first color component image separated by the color component separation unit. A pixel thinning unit that thins out the pixels interpolated by the interpolation unit; and a pixel combining unit that combines the pixels of the first, second, and third color components.

  According to this configuration, the video reproduction device can easily restore the high-definition video signal compressed by the video compression device by thinning out the interpolated pixels. Further, since the recording data of the video signal is not divided into a plurality of parts, the video reproducing apparatus does not need a timing synchronization circuit for reading a plurality of recording media or a plurality of recording areas in parallel, and the apparatus configuration can be simplified.

  The video compression method according to the present invention includes a pixel division step in which a control unit of a computer divides a pixel for each color component for a dual green single frame image having a Bayer array sampling structure, and in the pixel division step, A pixel interpolation step of interpolating the divided first color component image with an image obtained by converting each pixel of the image according to a geometric rule, the first color component after interpolation, and the An encoding step of performing encoding by a predetermined compression method is executed for each processing unit in which the second and third color components divided in the pixel dividing step are packaged.

  A video compression program according to the present invention includes a pixel division step for dividing a pixel for each color component in a dual-green single frame image having a Bayer sampling structure in a control unit of a computer, and the pixel division step. A pixel interpolation step of interpolating the divided first color component image with an image obtained by converting each pixel of the image according to a geometric rule, the first color component after interpolation, and the An encoding step of performing encoding by a predetermined compression method for each processing unit in which the second and third color components divided in the pixel dividing step are packaged.

  According to the present invention, a high-definition video signal can be efficiently encoded.

It is a figure which shows the function structure of the video recording / reproducing apparatus which concerns on 1st Embodiment. It is a figure which shows the function structure of the pre-processing part which concerns on 1st Embodiment. It is a figure which shows a mode that a single frame image is divided | segmented by the pixel division part which concerns on 1st Embodiment. It is a figure which illustrates the 1st pattern of the pixel interpolation process which concerns on 1st Embodiment. It is a figure which illustrates the 2nd pattern of the pixel interpolation process which concerns on 1st Embodiment. It is a figure which illustrates the 3rd pattern of the pixel interpolation process which concerns on 1st Embodiment. It is a figure which illustrates a DCT coefficient at the time of performing pixel interpolation concerning a 1st embodiment. It is a figure which illustrates the rearrangement process which concerns on 1st Embodiment. It is a figure which illustrates the packaging process which concerns on 1st Embodiment. It is a figure which shows the function structure of the post-processing part which concerns on 1st Embodiment. It is a figure which shows a mode that the DG signal which concerns on 1st Embodiment is decompress | restored. It is a figure which shows the function structure of the video compression transmission system which concerns on 2nd Embodiment. It is a figure which shows the relationship between PSNR obtained by simulation, and the file size after compression. It is a figure which shows the structure of a Bayer arrangement | sequence and a full resolution frame image. It is a figure which shows the processing procedure of JPEG. It is a figure which shows the technique of the compression recording by pixel division.

[First Embodiment]
The first embodiment of the present invention will be described below.
In the present embodiment, a case where JPEG is used as an encoding technique will be described as an example. However, the present invention can be similarly applied to encoding techniques using other spatial frequency transforms.

FIG. 1 is a diagram showing a functional configuration of a video recording / playback apparatus 1 (video compression apparatus, video playback apparatus) according to the present embodiment.
The video recording / reproducing apparatus 1 includes an input unit 11, a preprocessing unit 12, an encoding unit 13, a recording unit 14, a recording medium 15, a decoding unit 16, a postprocessing unit 17, and an output unit 18. Prepare.

The input unit 11 acquires a DG signal from an external device.
Examples of the external device include video equipment such as a CCU (Camera Control Unit), a recorder, and a switcher, in addition to a super high-definition single-panel camera and a four-panel camera having a DG signal interface. In addition, the external device may have a RAW signal output using a pixel arrangement similar to the DG signal in a high-definition image pickup device such as a 4K cinema camera or a digital still camera. In this case, the input unit 11 determines the format of the input signal and causes the preprocessing unit 12 to select an optimal processing method.

  The preprocessing unit 12 configures image data having a specified sampling structure, which is a processing unit of an encoding technique (for example, JPEG) used in the encoding unit 13, from the input DG signal.

FIG. 2 is a diagram illustrating a functional configuration of the preprocessing unit 12 according to the present embodiment.
The preprocessing unit 12 includes a blocking unit 121, a pixel dividing unit 122, a pixel interpolation unit 123, a rearrangement unit 124, and a packaging unit 125.

The blocking unit 121 is a circuit that buffers an input DG signal for each block of a predetermined size (for example, 8 × 8 pixels in the case of JPEG), and constitutes a processing unit of an encoding technique (for example, JPEG). is there.
When the processing unit of the encoding technique is 4 × 4, 16 × 16, 32 × 32 or the like, or when the processing unit is a rectangle such as 4 × 8, 8 × 16, etc., the blocking unit 121 may Blocking is performed according to the processing unit.

  The pixel dividing unit 122 is a circuit that divides the DG signal into R, G, and B. The divided G (G1 and G2) image data is supplied to the pixel interpolating unit 123, and the B image data and the R image data. Are respectively input to the rearrangement unit 124.

FIG. 3 is a diagram illustrating a state in which a single frame image of a DG signal is divided by the pixel dividing unit 122 according to the present embodiment.
The R, G1, G2, and B pixels arranged in the Bayer array include image data including missing pixels composed of G1 and G2, image data including missing pixels composed only of R, and missing pixels composed of only B. Divided into image data.

The pixel interpolating unit 123 interpolates the image of the G component (first color component) divided by the pixel dividing unit 122 by synthesizing it with an image obtained by converting each pixel of the image according to a geometric rule, This circuit generates image data without missing pixels.
As the geometric rule, for example, the following three patterns are adopted.

FIG. 4 is a diagram illustrating a first pattern of pixel interpolation processing according to this embodiment.
In this example, the pixel interpolation unit 123 inverts G1 and G2 of the DG signal in the horizontal direction for each block (8 × 8 pixels), and inserts it as an interpolation value for a lack region other than G1 and G2, thereby 8 × It is converted into a G signal without 8 missing pixels.

  The G signal obtained in this way has geometric symmetry in the horizontal direction in units of 8 × 8 blocks. Therefore, the DCT coefficients obtained by DCT are discrete in the vertical direction, and the entropy code in the subsequent stage. As a result, efficient compression becomes possible.

FIG. 5 is a diagram illustrating a second pattern of pixel interpolation processing according to this embodiment.
In this example, the pixel interpolating unit 123 inverts G1 and G2 of the DG signal in the vertical direction for each block (8 × 8 pixels), and inserts it as an interpolation value for a lack region other than G1 and G2, thereby 8 × It is converted into a G signal without 8 missing pixels.

  Since the G signal obtained in this way has an 8 × 8 block unit and geometric symmetry in the vertical direction, the DCT coefficient obtained by DCT becomes discrete in the horizontal direction, and the entropy code in the subsequent stage As a result, efficient compression becomes possible.

FIG. 6 is a diagram illustrating a third pattern of pixel interpolation processing according to this embodiment.
In this example, the pixel interpolation unit 123 rotates the G1 and G2 of the DG signal 90 degrees clockwise or counterclockwise around the center point for each block (8 × 8 pixels), and lacks other than G1 and G2. By inserting it as an interpolated value of the region, it is converted into a G signal without 8 × 8 missing pixels.

  Since the G signal thus obtained has a geometric feature in the diagonal direction in units of 8 × 8 blocks, the DCT coefficient obtained by DCT becomes discrete in the diagonal direction, and entropy coding is performed in the subsequent stage. Thus, efficient compression becomes possible.

FIG. 7 is a diagram illustrating DCT coefficients when pixel interpolation according to the present embodiment is performed.
(A) is a DCT coefficient when DCT is performed on an 8 × 8 pixel original image. In this case, large values are concentrated in the low frequency region, and values in the high frequency region are small.

  (B) is a DCT coefficient when the pixel interpolation processing of the first pattern (FIG. 4) is performed after sampling the G1 and G2 signals of the same original image as (a). In this case, coefficient values are generated discretely in a direction parallel to the v-axis, and columns including other than 0 and 0 columns appear alternately in the u-axis direction.

  (C) is a DCT coefficient when the pixel interpolation processing of the second pattern (FIG. 5) is performed after sampling the G1 and G2 signals of the same original image as (a). In this case, coefficient values are generated discretely in a direction parallel to the u axis, and rows including non-zeros and zero rows appear alternately in the v axis direction.

  (D) is a DCT coefficient when the pixel interpolation processing of the third pattern (FIG. 6) is performed after sampling the G1 and G2 signals of the same original image as (a). In this case, coefficient values are generated discretely on a diagonal line. In both the u-axis direction and the v-axis direction, 0 and a value including other than 0 appear alternately, and a row of 0 appears in the diagonal direction.

  When 0 appears regularly as in (b) to (d), it is advantageous for entropy coding, and the coding efficiency is improved.

  Note that the interpolation processing of the pixel interpolation unit 123 illustrated in FIGS. 4 to 6 may be performed on hardware using a transformation matrix filter. As a result, pixel interpolation can be realized without complicated calculations. In addition, the pixel interpolation unit 123 may realize pixel interpolation by software processing.

  The rearrangement unit 124 rearranges the R component (second color component) and B (third color component) images divided by the pixel division unit 122 into a structure in which missing pixels are thinned out.

FIG. 8 is a diagram illustrating a rearrangement process according to this embodiment.
In this example, the rearrangement unit 124 rearranges the R component and the B component to 4 × 4 for each 8 × 8 block in the DG signal.

  The packaging unit 125 packages the G signal obtained by the pixel interpolation unit 123 and the R signal and B signal obtained by the rearrangement unit 124 in units of MCU (Minimum Coded Unit) and supplies the packaged signals to the encoding unit 13. To do.

FIG. 9 is a diagram illustrating a packaging process according to this embodiment.
In this example, 4 blocks of 8 × 8 pixel G signals (16 × 16 pixels in total) and 4 blocks of 4 × 4 pixel R and B signals (8 × 8 pixels in total) are packaged as 4: 2 : Shows the case of handling as an MCU with a 0 sampling structure.

  The above description is the basic functional configuration of the preprocessing unit 12. In addition to this, a noise removal circuit, a color correction circuit, a color space conversion circuit such as YUV conversion, etc. May be provided.

Returning to FIG. 1, the encoding unit 13 compresses the data amount of the MCU supplied from the preprocessing unit 12 by encoding using a predetermined compression method. At this time, since the G signal is converted into a structure having a discrete two-dimensional histogram in the preprocessing unit 12, the compression efficiency of entropy coding can be improved.
In particular, when the encoding method is JPEG, the encoding unit 13 performs Huffman encoding by zigzag scanning the DCT coefficients. In the present embodiment, serial data after scanning is continuously 0, or 0 and an integer are alternately repeated. Therefore, the encoding unit 13 selects the optimal Huffman table, Compression efficiency can be increased.

The recording unit 14 converts the data encoded by the encoding unit 13 into a serial number file or a moving image file, wraps the data into a container format such as MXF (Material eXchange Format), and records the data on the recording medium 15.
At the time of video reproduction, the recording unit 14 reads the encoded recording data from the recording medium 15 and supplies it to the decoding unit 16.
Furthermore, the recording unit 14 may be provided with a function of reflecting, on the recording medium 15, the result of the user performing an operation such as changing the order of recording data or erasing from the operation panel or the like.

  The recording medium 15 is configured using a device capable of high-speed transfer, such as an SSD (Solid-State Drive), an HDD (Hard Disk Drive), or a magnetic tape. The recording medium 15 may perform parallel recording using a RAID configuration in order to increase capacity, increase transfer speed, improve reliability, and the like. In this case, the RAID control is preferably performed by the recording unit 14.

  The decoding unit 16 decodes the encoded data supplied from the recording unit 14 and supplies the decoded image data to the post-processing unit 17.

  The post-processing unit 17 converts the image data decoded by the decoding unit 16 and restores the DG signal input to the video recording / reproducing apparatus 1.

FIG. 10 is a diagram illustrating a functional configuration of the post-processing unit 17 according to the present embodiment.
The post-processing unit 17 includes a package dividing unit 171, a color component separating unit 172, a pixel thinning unit 173, a rearrangement unit 174, and a pixel composition unit 175.

  The package dividing unit 171 is a circuit that divides the signal supplied from the decoding unit 16 into MCU units and further blocks (8 × 8 pixels) and supplies the divided signals to the color component separating unit 172.

  The color component separation unit 172 is a circuit that separates the R, G, and B color components in the block. The separated G component is supplied to the pixel thinning unit 173, and the B component and R component are supplied to the rearrangement unit 174. input.

  The pixel thinning unit 173 thins out the pixels interpolated by the pixel interpolating unit 123 from the G component (first color component) image data separated by the color component separating unit 172 to obtain the G1 and G2 components of the original image. It is a circuit to take out only.

  The rearrangement unit 174 rearranges the pixels of the R component (second color component) and B component (third color component) images separated by the color component separation unit 172 to the missing pixel position of the G component. Circuit.

  The pixel combining unit 175 is a circuit that generates a DG signal by combining the G1 component and G2 component output from the pixel thinning unit 173 with the R component and B component output from the rearrangement unit 174.

FIG. 11 is a diagram illustrating how the DG signal is restored by the pixel thinning unit 173, the rearrangement unit 174, and the pixel synthesis unit 175 according to the present embodiment.
Thus, the G1 and G2 image data in which half of the pixels from the G component are thinned out and the image data in which the R component and the B component are rearranged in the gap between the G1 and G2 are combined, and the DG without missing pixels. A signal is generated.

  The output unit 18 outputs the DG signal obtained from the post-processing unit 17 to an external device. The external device may be a video processing device such as a color grading device or a transmission device in addition to a Super Hi-Vision monitor having a DG signal interface.

  According to the present embodiment, the video recording / playback apparatus 1 generates a full-resolution G signal by performing an interpolation process having a geometric feature on the G signal in the frame image of the Bayer array of the DG signal. . In addition, the video recording / reproducing apparatus 1 performs compression coding on the signal obtained by using such an interpolation process by an existing coding method having a sampling structure of 4: 2: 0.

  Therefore, since the video recording / reproducing apparatus 1 can compress the DG signal using the existing encoding method, the cost of the encoding hardware can be suppressed. Further, the video recording / reproducing apparatus 1 can perform compression while maintaining the continuity of the image without spatially dividing the frame image of the DG signal into a plurality of images, so that compression distortion can be suppressed. Furthermore, the video recording / reproducing apparatus 1 employs a pixel interpolation method having a geometric feature, so that a two-dimensional histogram after spatial frequency conversion is converted into a specific component (u-axis direction, v-axis direction or diagonal). Therefore, the compression efficiency by entropy encoding is improved, and a high-definition video signal can be encoded efficiently.

  At this time, since the video recording / reproducing apparatus 1 divides into processing block units of the encoding scheme and performs the interpolation processing for each block, the encoding efficiency can be improved by the interpolation processing adapted to the specified encoding scheme.

  Further, the video recording / reproducing apparatus 1 can easily create a geometric feature after the interpolation process by image conversion such as horizontal inversion, vertical direction inversion, or 90-degree rotation, and improve the encoding efficiency.

  In addition, when the video recording / playback apparatus 1 records a compressed image, the recording data is not divided into a plurality of parts, so that operations such as management of the recording medium, browsing of the recording data and backup are facilitated. Further, when reproducing recorded data, a timing synchronization circuit for reading out a plurality of recording media or a plurality of recording areas in parallel is not required, so that the apparatus configuration of the video recording / reproducing apparatus 1 can be simplified.

[Second Embodiment]
Hereinafter, a second embodiment of the present invention will be described.
In addition, about the structure similar to 1st Embodiment, the same code | symbol is attached | subjected and description is abbreviate | omitted or simplified.

FIG. 12 is a diagram showing a functional configuration of the video compression transmission system 2 according to the present embodiment.
The video compression transmission system 2 transmits the DG signal via the transmission path 3 from the transmission device 2a (video compression device) to the reception device 2b (video reproduction device).

The transmission device 2 a includes an input unit 11, a preprocessing unit 12, an encoding unit 13, a modulation unit 21, and a transmission unit 22.
The receiving device 2b includes a receiving unit 23, a demodulating unit 24, a decoding unit 16, a post-processing unit 17, and an output unit 18.
Here, the input unit 11, the pre-processing unit 12, the encoding unit 13, the decoding unit 16, the post-processing unit 17, and the output unit 18 have the same configuration as that of the video recording / reproducing apparatus 1 (FIG. 1) of the first embodiment. It's okay.

The modulation unit 21 converts the compressed encoded data into a transmission signal suitable for the transmission path 3.
The transmission unit 22 sends a transmission signal to the transmission path 3 as a physical signal.
The transmission path 3 may be an electric communication path, an optical fiber, a waveguide, a carrier wave in free space, or the like.

The receiving unit 23 receives the transmitted signal.
The demodulator 24 returns the received transmission signal to encoded data.

  According to the present embodiment, the video compression transmission system 2 can efficiently encode a high-definition video signal, and therefore can transmit video with high image quality even in a limited transmission band.

[Example]
Hereinafter, simulation procedures and results performed to show the effects of the above-described embodiment will be described.
In this example, a full-color high-definition image (resolution 1920 × 1080) is used as a test image to be recorded and reproduced, but the same result is obtained when a super high-definition image or another high-definition image is input. I can expect.

  FIG. 13 is a diagram showing the relationship between the PSNR (peak signal-to-noise ratio) obtained by the simulation according to the present embodiment and the file size after compression of the G signal.

Here, DCT localization 1 and DCT localization 2 in the graph show the results of the compression method of the embodiment, and are simulated by the following procedure.
(Procedure 1) A test image is subsampled into a Bayer array.
(Procedure 2) Among the images obtained in (1), pixel interpolation is performed on the G signal by the method of the first pattern (FIG. 4) and the third pattern (FIG. 6) in units of 8 × 8 blocks.
(Procedure 3) G-interpolated G signal is compression encoded by JPEG.
(Procedure 4) The file size of the G signal encoded in (3) is measured.
(Procedure 5) The encoded signal is decoded.
(Procedure 6) The image obtained in (5) is sub-sampled into a Bayer array.
(Procedure 7) The PSNR is calculated for the G signal obtained in (1) and (6).

Further, as a comparison object of the above method, a simulation result according to the following procedure is shown as a conventional method.
(Procedure 1) A test image is subsampled into a Bayer array.
(Procedure 2) The image obtained in (1) is divided into four low-resolution images G1, G2, R, and B.
(Procedure 3) The G1 and G2 signals of the image obtained in (2) are separately compressed and encoded by JPEG.
(Procedure 4) The file sizes of the G1 and G2 signals encoded in (3) are measured.
(Procedure 5) The encoded signal is decoded.
(Procedure 6) The PSNR is calculated for the G signal obtained in (1) and (5).

  In the simulation, a standard program (cjpeg / djpage) provided by IJG (Independent JPEG Group) was used as JPEG encoding software. In particular, the table generation for Huffman coding was optimized using cjpeg's optimize switch.

  When the method of the embodiment is used, the PSNR is improved by about 2 dB compared to the conventional method. In the method of the embodiment, considering that the file size of the uncompressed G signal is 2025 KB and is increased from 1012.5 KB of the conventional method, the method of the embodiment is higher compression than the conventional method. Nevertheless, it can be seen that the image quality is excellent.

  As mentioned above, although embodiment of this invention was described, this invention is not restricted to embodiment mentioned above. Further, the effects described in the present embodiment are merely a list of the most preferable effects resulting from the present invention, and the effects of the present invention are not limited to those described in the present embodiment.

As the frequency transform used in the encoding method, in addition to DCT, it can be widely applied to those that perform discrete sine transform (DST), Hadamard transform, wavelet transform, and other spatial frequency transforms.
In the present invention, a known interpolation method can be used in addition to the interpolation processing having geometric symmetry. At that time, the user may be able to switch between a plurality of interpolation methods.

In the above-described embodiment, the video recording / playback apparatus 1 having both the video recording function and the playback function has been described. However, the present invention is not limited to this.
For example, the decoding unit 16, the post-processing unit 17, and the output unit 18 may be separated as separate playback devices. Further, the recording unit 14 and the recording medium 15 may be separated as separate devices.

  In the present embodiment, the configuration and operation of the video recording / playback apparatus have been mainly described. However, the present invention is not limited to this, and includes each component, and is configured as a method or program for recording and playing back video. Also good.

  Further, the present invention may be realized by recording a program for realizing the function of the video recording / reproducing apparatus on a computer-readable recording medium, causing the computer system to read and execute the program recorded on the recording medium. Good.

  The “computer system” here includes an OS and hardware such as peripheral devices. The “computer-readable recording medium” refers to a storage device such as a portable medium such as a flexible disk, a magneto-optical disk, a ROM, and a CD-ROM, and a hard disk built in the computer system.

  Furthermore, “computer-readable recording medium” means that a program is dynamically held for a short time, like a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line. It is also possible to include one that holds a program for a certain time, such as a volatile memory inside a computer system that becomes a server or client in that case. Further, the program may be for realizing a part of the above-described functions, and may be capable of realizing the above-described functions in combination with a program already recorded in the computer system. .

1. Video recording / playback device (video compression device, video playback device)
2 Video compression transmission system 2a Transmission device (video compression device)
2b Receiver (video playback device)
DESCRIPTION OF SYMBOLS 11 Input part 12 Pre-processing part 13 Encoding part 14 Recording part 15 Recording medium 16 Decoding part 17 Post-processing part 18 Output part 21 Modulating part 22 Transmitting part 23 Receiving part 24 Demodulating part 121 Blocking part 122 Pixel dividing part 123 Pixel interpolation Unit 124 rearrangement unit 125 packaging unit 171 package division unit 172 color component separation unit 173 pixel thinning unit 174 rearrangement unit 175 pixel synthesis unit

Claims (8)

For a dual green single frame image having a Bayer array sampling structure, a pixel dividing unit that divides pixels for each color component;
An image of the first color component divided by the pixel division unit, and a pixel interpolating unit that interpolates the image and summing Rukoto moved all the deficient area according to the geometric rule each pixel of the image,
An encoding unit that performs encoding by a predetermined compression method for each processing unit that packages the first color component after interpolation and the second and third color components divided by the pixel dividing unit; Video compression device provided. A blocking unit that divides the single frame image into blocks suitable for the processing unit;
The video compression apparatus according to claim 1, wherein the pixel interpolation unit performs an interpolation process for each block. The pixel interpolation unit, the image of the first color component, the image compression apparatus according to claim 1 or claim 2, interpolated by summing Rukoto the image obtained by inverting the pixels of the image in the horizontal direction . The pixel interpolation unit, the image of the first color component, the image compression apparatus according to claim 1 or claim 2, interpolated by summing Rukoto the image obtained by inverting the pixels in the vertical direction of the image . The pixel interpolation unit, wherein the image of the first color component, to claim 1 or claim 2, interpolated by Rukoto alignment adding the image rotated 90 degrees around the center point of each pixel of the image Video compression device. The video compression apparatus according to any one of claims 1 to 5, wherein a decoding unit that decodes video data encoded by the predetermined compression method;
A color component separation unit that separates the first, second, and third color components for the image data decoded by the decoding unit;
A pixel thinning unit that thins out pixels interpolated by the pixel interpolation unit from the first color component image separated by the color component separation unit;
And a pixel combining unit that combines the pixels of the first, second, and third color components. The computer controller
For a dual green single frame image having a Bayer array sampling structure, a pixel dividing step for dividing pixels for each color component;
An image of the first color component is divided in the pixel dividing step, and the pixel interpolation step of interpolating the image and summing Rukoto moved all the deficient area according to the geometric rule each pixel of the image,
An encoding step of performing encoding by a predetermined compression method for each processing unit in which the first color component after interpolation and the second and third color components divided in the pixel dividing step are packaged; The video compression method to execute. In the control part of the computer,
For a dual green single frame image having a Bayer array sampling structure, a pixel dividing step for dividing pixels for each color component;
An image of the first color component is divided in the pixel dividing step, and the pixel interpolation step of interpolating the image and summing Rukoto moved all the deficient area according to the geometric rule each pixel of the image,
An encoding step of performing encoding by a predetermined compression method for each processing unit in which the first color component after interpolation and the second and third color components divided in the pixel dividing step are packaged; Video compression program to be executed.

Images