JP3149898B2 - Image information compression method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image information compression method for compressing moving image information of a high definition television (hereinafter, referred to as "HD-TV").
The present invention relates to an improvement in an image information compression method suitable for obtaining image information of a standard definition television (hereinafter, referred to as “SD-TV”) such as NTSC television from image information of a TV.

[0002]

2. Description of the Related Art Hierarchical high-efficiency coding has been studied for digital transmission and recording of HD-TV image information (Tsunashima et al., "HD-TV Hierarchy Considering Interlace Structure". Study on Coding ”, 199
3rd year, IEICE Spring Conference, D-253).
This is achieved by extracting and decoding a part of a bit stream obtained by encoding HD-TV image information, thereby achieving S
This is to obtain D-TV image information.
According to this, use of sub-band coding is considered as one of the coding methods.

[0003] For example, an HD-
Consider, as an example, hierarchical encoding of a TV image signal and an SD-TV image signal having an interlaced structure. First, HD-
The image signal of the TV1 field is divided into two band signals of a low band and a high band in the horizontal direction. Next, those band signals are divided into two band signals of a low band and a high band in the vertical direction. FIG. 4 shows this state, and the HD-TV
The image signal is sub-band divided into a total of four band signals LL, HL, LH, and HH. Then, the band signal LL having a low band in both the horizontal direction and the vertical direction is set as a video signal of the SD-TV1 field. Then, two SD-TV1 field signals are collected to form an SD-TV1 frame video signal.

[0004] By the way, in coding moving picture information such as a picture signal of a television, it is known that a code amount compression efficiency is improved by utilizing a correlation between frames. MPEG (Motion Picture Image Coding Expert
s Group) and MPEG2 also perform coding with high compression efficiency by a combination of inter-frame prediction and DCT (Watanabe, "MPEG2 Inter-frame Prediction Method", Technical Report of the Institute of Television Engineers of Japan, ITEJ Technical Report Vol.16, N
o.61, pp.37-42, ICS'92-73 (Oct.1992)).

[0005] From such a point, a method of compressing the HD-TV video signal divided into sub-bands using the correlation between frames as shown in FIG. 4 can be considered. That is, the horizontal and vertical low-band signals LL (or two-frame image signals of the SD-TV1 field) are output.
Signal from the field collection into one frame)
Inter-frame prediction may be performed in the same manner as in a conventional method such as MPEG or MPEG2. That is, the picture type of the frame is set to I picture (intra-coded screen), P
The picture is divided into a picture (inter-frame coding screen) and a B picture (bidirectional prediction / interpolation screen). For P pictures and B pictures, motion-compensated prediction coding is performed between an I picture or a P picture as a reference frame.

More specifically, the current frame (P or B picture) is divided into macroblocks of horizontal m pixels × vertical n pixels. On the other hand, an approximate image in the reference frame, that is, a block of m horizontal pixels × n vertical pixels for which motion compensation has been performed is selected. Then, it adaptively selects either the difference value for each pixel between the macroblock of the current frame and the block of the reference frame and the original pixel value of the macroblock in the current frame. That is, when there is no difference between the two images, a difference value is selected, and when there is a difference, an original pixel value is selected, and then the selected pixel value is encoded. However, the reference frame is also divided into sub-bands as shown in FIG. 4, and the block of horizontal m pixels × vertical n pixels is selected from the LL signals of the horizontal and vertical low bands.

The other bands HL, LH,
The motion compensation inter-frame prediction is similarly performed for HH. The reference frame is similarly divided into sub-bands, and a block of horizontal m pixels × vertical n pixels at a predetermined position is selected from the corresponding band signals of HL, LH, and HH to calculate a difference value.

[0008]

When such motion-compensated inter-frame prediction is performed, the coding amount of the current frame can be compressed with high efficiency for the LL signal in the horizontal and vertical low bands. This can also be explained by the fact that the statistical properties of the LL signal are similar to the statistical properties of the original image signal before subband division (Takishima et al., “Image quality evaluation method and statistical method of subband coding”). Study on Dynamic Properties ", IEICE Technical Report IE 91-4).

However, HL, LH,
With respect to the HH signal, the code amount of the current frame does not decrease so much as compared to before the motion compensation inter-frame prediction.
As a result, the code amount of the entire LL to HH image increases, and transmission and recording when transmitting and recording HD-TV image information are performed.
The recording capacity increases. Further, when the code amount of the entire image is limited to a certain fixed amount, it is necessary to reduce the code amount by requantization, and as a result, the image quality at the time of decoding deteriorates.

The present invention focuses on these points.
It is an object of the present invention to provide an image information compression method capable of reducing a code amount of a high-band signal in sub-band coding such as an HD-TV signal.

[0011]

In order to achieve the above object, the present invention subdivides a predetermined unit of image signal into a low band signal and a high band signal, and performs motion compensation prediction coding for each band signal. In the image information compression method for performing inter-frame encoding according to the above, for the high band signal, when there is no difference in the image signal between the pixel block of the current image and the motion compensated pixel block of the reference image , each of the pixels Block of
Select a difference value, and when there is a difference between the image signals,
When the pixel value of the pixel block of the current image is selected and encoded,
In addition to the selection, the signal between each of the pixel blocks
Is inverted, the sum of each pixel block
Can be selected and encoded .

[0012]

According to the present invention, in addition to the difference value between the image signals of the pixel block of the current image and the pixel block of the reference image, which are the targets of the adaptive selection conventionally performed, in addition to the original pixel value of the current image, , The sum of the image signals of these pixel blocks is also selected. For the high band component of the image signal,
The polarity may be reversed between the current image and the reference image. In such a case, encoding using a smaller amount of information can be performed by using the added value than the difference value.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the image information compression method according to the present invention will be described below in detail with reference to the accompanying drawings. <First Embodiment> First, a first embodiment of the present invention will be described with reference to FIGS. FIG. 1 shows a processing procedure of the first embodiment as a flowchart.
The processing shown in FIG. 1 is executed in a series of image information compression processing. As shown in FIG. 4, the HD-TV image signals of the current frame and the reference frame to be compressed are, as shown in FIG. The sub-band is divided into a band HL signal, a horizontal low band and a vertical high band LH signal, and a horizontal high band and a vertical high band HH signal (step S1).

Next, the current frame is divided into macroblocks of horizontal m pixels × vertical n pixels for each of the band signals LL, HL, LH, HH (step S2). Then, motion compensation prediction is performed between the same band signals of the current frame and the reference frame. [Case of LL band signal (case of Y in step S3)]

First, a block of m horizontal pixels × n vertical pixels of an image (motion compensated image) approximated from the LL signal of the reference frame is selected for each macro block of the LL band signal of the current frame. The displacement between the position of the macroblock in the current frame and the position of the block in the reference frame at this time is set as a motion vector (step S4). Then, the two blocks are compared, and one of the difference values for each pixel value of both blocks, the original pixel value of the macroblock of the current frame, is adaptively selected (step S5).

Next, one of the selected values is encoded as compression information of the macroblock (step S).
6, S7). Specifically, (a) when the above-mentioned difference value is selected, the macro block is set as an inter block (inter-frame coded block). A code indicating that the block is an inter block, a code indicating that a difference value is selected in the block (this code may be omitted), a code indicating a difference value for each pixel, and a position of a macro block And a code representing a motion vector representing an amount of deviation from the position of the block in the reference frame, as a compression code for the macroblock (step S6).

(B) When the original pixel value is selected, the macro block is set as an intra block (intra-frame coded block). Then, a code representing the intra block and a code representing the original pixel value are output as compression codes of the macro block (step S7). [In case of band signal other than LL (N in step S3)]

Next, for the HL, LH, and HH band signals other than the LL signal, an image (movement) approximated from the reference frame in macroblock units using the motion vector obtained for the corresponding macroblock of the LL signal. A block of horizontal m pixels × vertical n pixels of the compensation image) is selected and compared (steps S8 and S9).

Then, one of the difference values for each pixel value of both blocks, the added value for each pixel value of both blocks, and the original pixel value of the macroblock in the current frame is adaptively selected (step S9). It goes without saying that this selection is the same for all pixels in the block. That is, in the prior art, for the high band signal, any of the original pixel values with the same difference value as the low band signal is adaptively selected.
In this embodiment, the added value is also a selection target.

Next, one of the selected values is encoded as compression information of the macroblock (step S).
10, S11, S12). Specifically, (c) when the above difference value is selected, the macro block is set as an inter block. Then, a code indicating that the block is an inter block, a code indicating that a difference value is selected in the block, a code indicating a difference value for each pixel, a position of a macroblock and a position of a block in a reference frame, Is output as the compression code of the macroblock (step S10).

(D) When the added value is selected, the macro block is set as an inter block, a code indicating that the block is an inter block, a code indicating that the added value is selected in the block, A code representing an added value for each pixel and a code representing a motion vector representing a shift amount between the position of the macroblock and the position of the block in the reference frame are output as compression codes of the macroblock (step S11).

(E) When the original pixel value is selected, the macro block is set as an intra block. Then, a code representing the intra block and a code representing the original pixel value are output as compression codes of the macro block (step S12).

As described above, according to the present embodiment, at the time of motion-compensated inter-frame prediction, for a high-frequency signal other than the LL signal, an addition value can be selected as an encoding target in addition to the conventional difference value. It is characterized by having. As for the other parts, known techniques such as MPEG and MPEG2 may be appropriately combined as needed.

Next, the operation of this embodiment will be described with reference to FIG. In order to facilitate understanding, FIG.
A one-dimensional image model in the horizontal direction as shown in FIG. In the figure, the horizontal axis represents the horizontal position Y of the pixel, and the vertical axis represents the signal level L of the pixel. A vertical line indicates a digitized pixel signal, and a dotted line indicates a corresponding analog signal. When such an image signal is separated into a low-band signal and a high-band signal, the signals are as shown in FIGS. The low band signal corresponds to the LL signal in FIG. 4, and the high band component corresponds to the HL signal. In each case, the sampling frequency is １ ／ compared to the signal of FIG.

Here, it is assumed that the reference frame is an image slightly moved with respect to the current frame, as shown in FIG. The image signal of the reference frame shown in FIG.
Similarly, when the signal is separated into a low-band signal and a high-band signal, the signals are as shown in FIGS.

First, a motion vector is obtained for the low-band signal in the current frame and the reference frame. Here, the macro block is assumed to be four pixels indicated by WA in FIG. As shown in FIGS. 3C and 3D, Δ
Assuming that a motion vector having a size of? Is obtained, motion compensation prediction is performed using four pixels indicated by WB in FIG. At this time, if the difference value between the two is encoded as in the related art, sufficiently satisfactory encoding is possible.
For example, when focusing on the pixel PA, the level is L1, and the level of the motion-compensated pixel PB is L2. Therefore, the difference value L1-L2 is encoded.

On the other hand, when comparing the current frame and the reference frame for the high band signal, FIG.
As shown in (2), even if the motion is compensated, the sign of the signal level may be reversed depending on the pixel. This is shown in FIG.
(F) is to reduce the sampling frequency to １ ／ with respect to (A) and (B). Before と す る, the pixel whose sign is not reversed as shown by the dotted line in the figure. It is possible to choose. For example, focusing on the pixel PC,
The level is L3 in the current frame of FIG.
The level of the pixel PD motion-compensated in the reference frame is L4
Since ≒ −L3, the difference value is L3−L4 ≒ L3 + L3
Will be. For this reason, even if the difference value is encoded with respect to the high-band signal as in the related art, the information amount is not so compressed.

Therefore, in this embodiment, the pixels PC, P
Considering that the sign of the signal level may be inverted as in D, the added value is used as described above. That is, when an added value is obtained for the PC, L3 + L4 = L3-L3 ≒ 0. Therefore, if this is encoded, the code amount is reduced.

In FIG. 2, the motion vector obtained from the low-band signal is used as it is as the motion vector of the high-band signal, but a search may be made around the motion vector obtained from the low-band signal. At this time, the smaller of the absolute value sum of the difference values for each pixel in the block and the sum of the absolute values of the added values for each pixel value in the block, etc. It is natural from the above point that the value of may be used.

As described above, according to the present embodiment, the HL signal, L
The code amount of the H signal and the HH signal can be reduced.
As a result, the code amount of the entire image can be satisfactorily compressed, and the transmission capacity and recording capacity when transmitting or recording image information can be reduced. When the code amount of the entire image is limited to a certain fixed amount, the image quality at the time of decoding is improved.

<Second Embodiment> Next, a second embodiment of the present invention will be described with reference to FIG. In the first embodiment, the image signal is equally divided into four, namely, LL, HL, LH, and HH. However, the image signal may be further divided and then subjected to motion compensation prediction coding (Yasawa et al., " Frequency Domain Realization of Motion Compensated Interframe Prediction in Non-Equal Partition Subband / Wavelet Transform Coding ", IEICE IE91-83).

According to this, the LL signal, which is divided into a low band and a high band in each of the horizontal direction and the vertical direction, and the LL signal having a low band in both the horizontal direction and the vertical direction is motion-compensated for each recursively divided band. A method of performing inter-frame predictive coding has been pointed out. FIG. 3 shows the state of this band division. The divided signals shown in FIG. 4 are represented by LL0, HL0,
They are represented by LH0 and HH0. According to the second embodiment, the portion of the LL0 signal has a lower band in the horizontal and vertical directions.
It is divided into high band LL1, HL1, LH1, and HH1. Further, the portion of the LL1 signal is further divided in horizontal and vertical directions into LL2, HL2, LH2, and HH2 of a low band and a high band.

In such a case, the same processing as in the first embodiment is performed for signals in the low band both in the horizontal and vertical directions, that is, signals in other bands except for the LL2 signal in the lowest band. I do. That is, a block of horizontal m pixels × vertical n pixels at a certain position of the reference frame is compared with a macroblock in the current frame, and a difference value for each pixel value is calculated.
Either the added value for each pixel value or the original pixel value of the macroblock is adaptively selected and encoded. At this time, the size of the macroblock is changed for each band, as described in the above-mentioned document.

<Other Embodiments> The present invention is not limited to the above-described embodiments, but includes, for example, the following. (1) In the above-described embodiment, the motion vector with respect to the reference frame is obtained in units of integer pixels. In this case, an interpolated value between pixels is used as a motion compensation image. (2) In the above embodiment, a block of horizontal m pixels × vertical n pixels is used as a unit, but this is further divided into small blocks,
A difference value, an addition value, and the like may be adaptively selected for each of these small blocks. In this case, a sign indicating a difference value, an addition value, or the like is added to each small block.

(3) As the reference frame when the current frame is a B picture, any one of a preceding frame, a following frame, and an interpolated value obtained from the preceding and following frames can be selected. (4) In the above embodiment, motion compensation is performed between frames, but field / frame adaptive motion compensation may be performed. (5) The difference value, the added value,
DCT may be simply applied to any of the original pixel values, or the field / frame adaptive DC may be applied.
T may be applied.

(6) Various processes such as sub-band coding may be performed on each band signal of the current frame predicted by the motion compensated inter-frame prediction according to the above embodiment, if necessary. . (7) As one preferable application example of the present invention, a code is recorded on a medium for each sub-band shown in FIG. 4 and only the code of the LL signal is read out of the medium to read, for example, an NTSC video. In some cases, reproduction is performed, the entire code is read, and, for example, high-definition video reproduction is performed. But,
Subband coding In general, the invention is applicable.

[0037]

As described above, according to the image information compression method according to the present invention, when coding by motion compensation is performed for each sub-band divided band signal, the high band signal is referred to as the current image. Since the addition value between the comparison blocks of the image is also selected as the encoding target, there is an effect that the code amount of the high band signal can be reduced.

[Brief description of the drawings]

FIG. 1 is a flowchart showing a first embodiment of a method for compressing image information according to the present invention.

FIG. 2 is an explanatory diagram showing the operation of the first embodiment.

FIG. 3 is an explanatory diagram showing a state of subband division according to a second embodiment of the present invention.

FIG. 4 is an explanatory diagram illustrating an example of subband division of an HD-TV signal.

[Explanation of symbols]

LL, HL, LH, HH, LL0, HL0, LH0, HH
0, LL1, HL1, LH1, HH1, LL2, HL2, LH
2, HH2... Band signals divided into subbands, L1 to L4.
Signal level, PA, PB, PC, PD: pixel, WA: macroblock, WB: block from which motion vector was obtained,
Δ: The movement between the current frame and the reference frame obtained in a certain macroblock having a low-band signal.

──────────────────────────────────────────────────続 き Continuation of the front page (56) References JP-A-5-91498 (JP, A) JP-A-5-219490 (JP, A) JP-A-4-322593 (JP, A) 191800 (JP, A) JP-A-5-328326 (JP, A) JP-A-6-86254 (JP, A) JP-A-6-105300 (JP, A) JP-A-6-217291 (JP, A) JP-A-6-233283 (JP, A) IEICE Technical Report IE91-82 PCSJ'91, 4-1 1992 IEICE Spring Conference Proceedings D-302 Mitsubishi Electric Technical Report, Vol. 67 No. 7, p. 39-43 (58) Fields investigated (Int. Cl. ⁷ , DB name) H04N ^7/ 24-7/68 H04N 7/01

Claims (1)

(57) [Claims] An image information compression method for dividing an image signal of an image of a predetermined unit into a low-band signal and a high-band signal by sub-band, and performing inter-frame coding by motion compensation prediction coding for each band signal, For the high band signal, the image signal between the pixel block of the current image and the pixel block of
If there is no difference, select the difference value of each pixel block
And when there is a difference between the image signals, the pixels of the current image
While selecting and encoding the pixel values of the block,
In addition, the signal between the pixel blocks is inverted.
In this case, select the added value of each pixel block and
An image information compression method characterized in that it can be encoded.