CA2068073C - Encoding of motion picture signal - Google Patents
Encoding of motion picture signalInfo
- Publication number
- CA2068073C CA2068073C CA 2068073 CA2068073A CA2068073C CA 2068073 C CA2068073 C CA 2068073C CA 2068073 CA2068073 CA 2068073 CA 2068073 A CA2068073 A CA 2068073A CA 2068073 C CA2068073 C CA 2068073C
- Authority
- CA
- Canada
- Prior art keywords
- block
- valid
- blocks
- map
- valid block
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/90—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using coding techniques not provided for in groups H04N19/10-H04N19/85, e.g. fractals
- H04N19/98—Adaptive-dynamic-range coding [ADRC]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/503—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/20—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using video object coding
Abstract
A valid/invalid detection section 1 generates an original valid block map based on a difference between frames. The current and previous valid block maps are weighted within weight sections 2 and 3 respectively and are thereafter summed up to compose them. After isolated valid blocks are removed from the composed valid block map within a segmentation section 5, it is determined whether or not a moving object exists and the direction of the moving object is determined based on the distribution of the valid block map. If a significant movement is detected, then blocks around the object are weighted at the movement direction side. Another segmentation section 7 determines valid blocks with reference to interest blocks and neighbor blocks, and isolated invalid block section 8 removes isolated invalid blocks to generate a complete valid block map. Encoding is performed based on the complete valid block map.
Description
2~807~
. .
ENCODING OF MOTI~N PICTURE SIG~IAL
BACKGROUND OF THE INVENTION
The present invention relates to encoding a motion picture signal. using band compression techni~ues, and more particularly to detecting an interest image region, for example the image of a speaker using a video telephone, and then encoding only the interest image region.
As a prior approach for encoding a motion picture signal using band compression techni~ues, "A Color Motion Videophone for the ISDN", Report No.D-233, 1989 Spring Grand Conference, The Institute of Electronics, Information and Communication Engineers of Japan is known. According to this approach, ~ facial region is detected to generate a map representing the facial region9 and an image encoding section performs interframe/ inframe adaptive predictive encoding of picture elements of the current frame9 using picture elements of the previous frame and current adjacent picture elements. When a picture element to be encod~d is in the facial region~ the encoding is repeated to the final stage, and otherwise the encoding is stopped at the stage immediately before the final stage.
~owever, the prior approach still performs course encoding 2~g8~7~
of a background part or a part other than facial re~ion, and then noise related to the background part causes unnecessary information. Further when picture elements is converted from the background part to the facial region between consecutive frames, the course encoding is switched to the fine encoding, and then considerable predictive error signals are generated, which result in unnecessary information much more. Accordingly the encoding efficiency deteriorates.
Another prior approach is disclosed in "A method for facial region detection on a color video phone", Report No.D-92, 1989 Spring Grand Conference, The Institute of Electronics, Information and Communication Engineers of Japan. According to the prior approach, a facial region is detected using histograms of picture elements having values larger than a threshold value in a differential image between consecutive frames. At first a vertical histogram is generated by counting significant picture elements in the differential ~ image horizontally, and used for determination of the top of the face. And the image is divided into horizontal band sub-areas and then horizontal histograms are generated for the sub-areas by counting significant picture elements in the sub-areas vertically. The width of the ~ace is determined based on the several horizontal histograms for the sub-areas beneath the top position of the face, and height of the face is determined proportional to the face width.
2~g~
However this approach doesn t directly detect the outline of the face, and doesn t extract correctly the facial region for fine encoding. Further this article doesn t suygest encoding only the facial part to reduce unnecessary information caused by noise in the background part.
SUMMARY OF THE INVENTION
As with video telephones, while a speaker is in motion, the background part is still. According to the present invention, a moving interest part such as the face of a speaker is detected by using differential image between frames. Further th.~s detection is corrected based on the movement direction of the interest part. Responsive to the corrected detection, only a part of a motion picture signal corresponding to the interest part is exactly extracted and encoded. As a result, unnecessary encoding information generated by noise in the background can be obviated and the efficiency can be improved.
How to extract an interest part for example a speaker part is described with reference to the drawings. Suppose a speaXer moves as shown in FIG.2 from time "t0" to time "tl", and from time "tl" to time "t2". At first, regions having difference between the frames at the times tl and t2 are detected as illustrated in FIG.3 with hatching. Hatching portions in the background are caused by differential 2~)~81~rl3 signals generated from noise in the background. Ne~.t, the ~rame is divided into blocks of picture elements in m rows x n columns, and each block is determined as a valid if the sum of the absolute values o~ differential signals in the block is larger than or equal to a predetermined first threshold value, and the block is determined as invalid if the sum is less than the threshold value. A valid block map thus obtained is illustrated in FIG.4B, in which valid blocks are indicated with hatching.
Now suppose a valid block map obtained for frames at times tO and tl is illustrated in FIG.4A an~ the block is hereinafter referred to as a fifth valid block map or complete valid block map. And while the valid block map for the current frame, or the first valid block value is weighted With a first weight value, the valid block for the previous frame or the fifth valid block map is weighted with a second weight value. An example of the weighting is described as follows.
Suppose for example the weight value for valid blocks of the previous frame is "1" and that for its invalid blocks is "0", and the weight value for valid blocks of the current frame is "2" and that for its invalid blocks is "0" like that for the invalid blocks of the previous blocks. After weighting the valid block maps of the current and previous ~rames as mentioned above, the valid block maps are summed up to obtain a second valid block map. The second valid . .
:
: , , :
, 20~g~7~
block map is illustrated in EIG.5.
Next, segmentation is performed based on the composed valid block map. An example of the segmentation is described with reference to EIG's 5, 6, 7, and 8. In the valid block map, each block to be processed is set as valid if the sum of the values of the block and the eight neighbor blocks, that is, a block "k" and blocks "a", "b", "c", "d", "e", "f", "g", and "h" in FIG.7, is larger than or equal to a predetermined threshold value, and the block to be processed is set as invalid if the sum value is less than the threshold value.
A third valid block map is accordingly obtained as illustrated in FIG's.6A and 6B.
However, the above segmentation which uses the same threshold for all the blocks may make some blocks in the moving direction side area to b~ set as valid, changed to invalid and some portions of the person figure is dropped off as shown in FIG.6B, and an encoded:image~is degraded.
.
For avoiding that deterioration, values of the most left side valid blocks along horizontal block lines in the third valid block map are summed up, and likewise values of the most right side valid blocks along the horizontal block lines in the map are summed up. The difference between the left side valid block sum value and the right side valid block sum value is then calculated, and if the absolute 213~8~7~
value of the difference is larger than a predetermined third threshold value, then the valid blocks on the side having the larger sum value are weighted with a third weight value to avoid the valid area from dropping off.
As seen from the illustration of EIG.6A, the sum of the values of the most left side valid blocks along the horizontal block lines is "6" and the sum of the values of the most right side valid blocks along the horizontal block lines is "13". Suppose the absolute value of the dif~erence between the left and right side sums "7" (=13-6) is larger than the threshold. Then, the valid blocks on the right side are weighted with the third weight. In this case, the most right side valid blocks along the horizontal block lines are added by one and another valid block map is obtained as illustrated in FIG.8A.
After the weighting with the third weight as shown in FIG.
8A, each block in the third valid block map is set as valid with reference to the neighbor blocks if the sum value o~
the block and the neighbor blocks is larger than or equal to a predetermined fourth threshold value, and the block is set as invalid if the sum is less than the fourth threshold value. As a result, a fourth valid block map is obtained as - illustrated in FIG.8B in which valid areas are indicated with hatching.
If the absolute value of the difference between the sums of 2~G8~7~
the values of the blocks on the left and right sides is less than the third threshold value, there is no significant mGvement direction. Accordingly, the weighting with the third weight is not performed. And each block in the third valid block ~ap is set as valid wi-th reference to the neiyhbor blocks if the sum value of the block and the neighbor blocks is larger than or equal to a predetermined fifth threshold value, and the block is set as invalid if the sum is less than the fifth threshold value. As a result, the fourth valid block map is obtained.
As shown in FIG.8B, there can be isolated invalid blocks in the valid region. If ancoding is applied to only valid blocks in the fourth valid block map, the isolated invalid blocks in the valid region are not encoded, and an image of the previous frame remains at the isolated invalid blocks and causes discontinuation with the surrounding blocks.
This may result in significant deterioration of the encoded image. For resolving the issue, the same processing is applied to the isolated invalid blocks as used for the segmentation above mentioned. More particularly, in the fourth valid block map, each invalid block is checked with reference to the neighbor blocks, and if the sum of the values of the neighbor blocks is larger than or equal to a predetermined sixth threshold value, then the invalid block is filled with a value indicating validity. By this operation, the area of isolated invalid blocks in FIG.8B is removed, and a fifth valid block map is obtained. FIG.9 illustrate the fifth valid block map wherein the valid blocks are indicated with hatching.
Thereafter, the area of valid blocks or the speaker area of FIG.9 is encoded by using interframe and/or inframe correlations, and unnecessary information caused by the noise in the background is accordingly reliably removed. Therefore coding efflciency is improved.
The above threshold and weightlng values are determlned based on statlstical experiments. The arrangement of reference picture elements for the segmentatlon and the isolated block removal can be varled from the above.
In accordance with the present inventlon, there is provided a motlon plcture signal encoding method utllizing a correlation between images, comprising the following steps of:
generating a first valid block map for each frame, by dividing images of said frame and corresponding reference frame into blocks of a plurality of picture elements, detecting for each of said blocks difference between said images of said frame and said reference frame, making each of said blocks a valid block if said dlfference of sald block is larger than or equal to a predetermined first threshold value, and making said block an invalid block if said difference of said block is less than said first threshold value;
generating a second valid block map for each of said frames, by weighting said first valid block map with a first weighting factor, weighting a complete valld block map generated for a previous frame with a second weighting factor, and composing said weighted first and complete valid block 747~0-1 maps by sum operations;
generating a third valid block map for each of sald frames, by making each of said valid block of said second valid block map an invalid block if a sum of values of said valid block and neighbor blocks is less than a predetermined second threshold value;
generating a fourth valid block map for each of said frames, by, obtaining a sum of values of valid blocks at the most left locations along horizontal block lines in said third valid block map, obtaining a sum of values of valid blocks at the most right locations along horizontal block lines in said third valid block map, if a difference between said sums is larger than or equal to a third threshold value then weighting valid blocks at the most left or rlght locations having the greater sum with a third weighting factor and further maklng each block of sald thlrd valld block map thus weighted a valld block provided that a sum of values of sald block and neighbor blocks is larger than or equal to a fourth threshold value and making each of said block of said third valid block map thus weighted an invalid block provided that a sum of values of said block and neighbor blocks is less than said fourth threshold value to generate said fourth valid block map;
if said difference between said sums is less than said third threshold value then making each block of said third valid block map a valid block provided that a sum of values of -8a-747gO-l said block and neighbor blocks is larger than or equal to a fifth threshold value and making each of sald block of said third valid block map an invalid block provided that a sum of values of said block and nelghbor blocks is less than said fifth threshold value to generate said fourth valid block map;
generating said complete valid block map for each of sald frames, by making each of invalid blocks of said fourth valid block map a valid block if a sum of values of neighbor blocks is larger than or equal to a predetermined sixth threshold ~0 value; and encoding only a part of said image of said frame corresponding to valid blocks of said complete valid block map by using correlations between images and/or in a image.
In accordance with another aspect of the invention, there is provided a motion picture signal encoding apparatus utilizing a correlation between images, comprising:
means for generating a first valid block map for each frame, by dividing images of said frame and corresponding reference frame into blocks of a plurality of picture elements, detecting for each of said blocks difference between said images of said frame and said reference frame, making each of said blocks a valid block if said difference of said block is larger than or equal to a predetermined first threshold value, and making said block an invalid block if said difference of said block is less than said first threshold value;
means for generating a second valid block map for each of said frames, by weighting said first valid block map with a -8b-first weightinq factor, weighting a complete valid block map generated for a previous frame with a second factor, and composlng said weighted first and complete valid block maps by sum operations;
means for generating a third valid block map for each of said frames, by making each of said valld blocks of said second valid block map an invalid block if a sum of values of saidvalid block and neighbor blocks is less than a predetermined second threshold value;
means for generating a fourth valid block map for each of said frames, by, obtaining a sum of values of valid blocks at the most left locations along horizontal block lines in sald third valid block map, obtaining a sum of values of valid blocks at the most right locations along horizontal block lines in sald third valid block map, if a difference between said sums is larger than or equal to a third threshold value then welghting valid blocks at the most left or right locatlons havlng the greater sum with a third weighting factor and further making each block of said third valid block map thus weighted a valid block provided that a sum of values of said block and nelghbor blocks is larger than or equal to a fourth threshold value and making each of said blocks of said third valid block map thus weighted an invalid block provided that a sum of values of said block and neighbor blocks is less than said fourth threshold value to generate said fourth valid block map;
if said difference between said sums is less than said -8c-thlrd threshold value then making each block of said third valid block map a valid block provlded that a sum of values of sald block and neighbor blocks is larger than or equal to a fifth threshold value and making each of said blocks of said third valid block map an lnvalld block provlded that a sum of values of sald block and nelghbor blocks is less than said fifth threshold value to generate said fourth valid block map;
means for generatlng sald complete valld block map for each of said frames, by maklng each of invalid blocks of said fourth valld block map a valld block lf a sum of values of nelghbor blocks ls larger than or equal to a predetermined slxth threshold value; and means for encoding only a part of said lmage of sald frame correspondlng to valld blocks of sald complete valld block map by using correlatlons between lmages and/or ln an image.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 ls a block diagram illustrating an embodiment of the present inventlon.
FIG. 2 ls a drawing for lllustratlng the prlnclples of the present invention, in which a speaker is shown as moving from a tlme "t0" to "tl", from the time "tl" to "t2".
FIG. 3 is a drawlng for illustrating the principles of the present inventlon, whlch shows areas having a significant dlfference between the frames at the times "tl"
and "t2".
FIG.4 ls a drawlng for lllustratlng the prlnclples of the -8d-~g~
present invention, which shows a first valid block map (B) and a fifth valid block map of the previous frame (A).
FIG.5 is a drawing for illustrating the principles of the present invention, which shows a second valid block map.
FIG.6 is a drawing for illustrating the principles of the present invention, which shows a third valid block map (A), and an example incomplete encoding image caused by the third valid block map (B).
FIG.7 is a drawing for illustrating the principles of the present invention, which shows a block to be processed and neighbor blocks for segmentation.
FIG.8 is a drawing for illustrating the principles of the present invention, which shows a fourth valid block map (A), and the valid area in the fourth yalid block map (B).
FIG.9 is a drawing for illustrating the principles of the present invention, which shows the valid area in a fifth valid block map.
DETAILED DESCRIPTION OF A PREFERRED EMBODIMENT
One embodiment of the present invention is now described with reference to the drawings.
2~8~7~
FIG.1 shows the embodiment, in which an inputted motion picture signal is fed to a valid/invalid detection section 1 and a delay element 10 via a line 11. The valid/invalid detection section 1 stores a motion picture signal of a previous frame, and acquires the frame differential signal between the motion picture signal of the previous frame and that newly inputted via the line 11, and divides the frame differential signal into blocks of picture elements in m rows and n columns, and then calculates for each block the absolute value of the sum of frame differential values in the block. The valid/invalid detection section 1 further labels each block with "valid block" if the absolute value of the sum of the frame differential values in the block is larger than or equal to a predetermined first threshold value, and labels the block with "invalid block" if the absolute value of the sum of the differential values is less than the first threshold value, and accordingly obtains a first valid block map. The absolute value of the sum of the differential values can be replaced with a sum of squares, or the number of picture elements having values exceeding a threshold value.
The valid block signal obtained through the valid/invalid detection section 1 is fed to a weight section 2 which in turn per~orms a predetermined first weightlng operation on valid blocks in the valid block map fed from the valid/invalid detection section 1 to modify the values of the valid blocks. The first valid block map weighted within ,, ~ ,, , ~8~73 the weiyht section 2 is fed to a sum section 4 which sums the first valid block map thus fed and a fifth valid block map to generate a second valid block map thus weighted. The second valid block map obtained in the sum section 4 is fed to a segmentation section 5.
The segmentation section 5 calculates for every block in the second valid block map the sum of the values of the block and neighbor blocks, and makes the block an invalid block if its sum is less than a predetermined third threshold value to obtain a third valid block map.
The values in the third valid block map fed from the segmentation section 5 are checked along each horizontal block line in the map , from the left to the right, and the value of the first coming valid block is determined as the value at the left edge of the block line. Likewise, the values in the block map are checked along each horizontal block line, from the right to the left, and the value of the ! first coming valid block is determined as the value of the right edge of the block line. Thereafter, the sum of the values o~ the valid block on the left edges and that on the right edges are obtained, and the difference between the sums of the values of the valid blocks on the left edges and right edges is then calculated. If the absolute value of the difference is larger than or equal to a predetermined third threshold, then the valid blocks on the edge having a larger sum are weighted in a weight section 6 by addition of 2 ~ 7 3 a predetermined value.
The third valid block map which has undergone the third weighting operation is fed to a another segmentation section 7 in which each block is made a valid block if the sum of the values of the block and neighbor blocks is larger than or equal to a predetermined fourth threshold and the block is made an invalid block if the sum is less than the fourth threshold to generates a fourth valid block map.
When the difference between the sums of values of the valid blocks on the left and right sides is less than the third threshold value, in the segmentation section 7, each block is made a valid block if the sum of the values of the block and neighbor blocks is larger than or equal to a predetermined fifth threshold and the block is made an invalid block if the sum is less than the fifth threshold to generates the fourth valid block map.
As illustrated in FIG. 6, suppose that a block "k" is a block to be processed and that neighbor blocks are for example blocks "a", "b", "c", "d", "e", "~", "g", and "h".
If the sum of the their values is larger than or equal to the predetermined threshold value, then the block "k" is made a valid block, and the sum is less that the threshold value, then the block "k" is made an invalid block.
The fourth valid block map obtained through the segmentation ~g~7~
section 7 is supplied to an isolated invalid block removal section 8. The removal block section 8 removes isolated invalid blocks or invalid blocks in a valid block area to connect valid blocks. This process is performed with reference to neighbor blocks like the segmentation, by changing an invalid block to be processed into a valid block if the sum of values of the block and neighbor blocks is larger than or equal to a predetermined sixth threshold value. Thus, a fifth valid block map in which isolated invalid blocks have been removed is obtained.
The fifth valid block map thus obtained is supplied to another weight section 3 which performs a predetermined second weighting operation on valid blocks in the fifth valid block map to modify the values of the valid blocks, and supplies the modified fifth valid block map to the sum section 4 upon the next frame processing.
The delay section lO compensates an inputted motion picture signal for the time delay from the time when the motion picture signal is inputted to the time the fifth valid block map is supplied to an encoding section 9, and arraigns the timing of the fifth valid block map and the inputted signal.
The time compensated signal outputted from the delay section 10 is fed to the encoding section 9.
The encoding section 9 does encode only a part of the motion picture signal inputted from the delay section 10 ~, .
~6~7~
corresponding to the valid blocks of the fifth valid block map fed from the isolated invalid block removal section 8, that is, corresponding to a portion indicated as a speaker area, but doesn t encode the background part indicated with the invalid blocks. For the encoding, methods utilizing a correlation between frames, such as motion compensation, methods utilizing a correlation in a frame such as orthoganal conversion, and methods utiliziny correlations between frames and/or in a frame can be used.
The above threshold values are determined according to values statistically obtained beforehand.
As seen from the above detailed description of the preferred embodiment, according to the present invention, only a speaker area obtained by segmentation is encoded, and unnecessary information caused by noise in the background is removed, and the encoding efficiency is accordingly irnproved.
Although the embodiment of the present invention have been described, it is und0rstood that numerous variations may be made in accordance with the principles of the invention.
. .
ENCODING OF MOTI~N PICTURE SIG~IAL
BACKGROUND OF THE INVENTION
The present invention relates to encoding a motion picture signal. using band compression techni~ues, and more particularly to detecting an interest image region, for example the image of a speaker using a video telephone, and then encoding only the interest image region.
As a prior approach for encoding a motion picture signal using band compression techni~ues, "A Color Motion Videophone for the ISDN", Report No.D-233, 1989 Spring Grand Conference, The Institute of Electronics, Information and Communication Engineers of Japan is known. According to this approach, ~ facial region is detected to generate a map representing the facial region9 and an image encoding section performs interframe/ inframe adaptive predictive encoding of picture elements of the current frame9 using picture elements of the previous frame and current adjacent picture elements. When a picture element to be encod~d is in the facial region~ the encoding is repeated to the final stage, and otherwise the encoding is stopped at the stage immediately before the final stage.
~owever, the prior approach still performs course encoding 2~g8~7~
of a background part or a part other than facial re~ion, and then noise related to the background part causes unnecessary information. Further when picture elements is converted from the background part to the facial region between consecutive frames, the course encoding is switched to the fine encoding, and then considerable predictive error signals are generated, which result in unnecessary information much more. Accordingly the encoding efficiency deteriorates.
Another prior approach is disclosed in "A method for facial region detection on a color video phone", Report No.D-92, 1989 Spring Grand Conference, The Institute of Electronics, Information and Communication Engineers of Japan. According to the prior approach, a facial region is detected using histograms of picture elements having values larger than a threshold value in a differential image between consecutive frames. At first a vertical histogram is generated by counting significant picture elements in the differential ~ image horizontally, and used for determination of the top of the face. And the image is divided into horizontal band sub-areas and then horizontal histograms are generated for the sub-areas by counting significant picture elements in the sub-areas vertically. The width of the ~ace is determined based on the several horizontal histograms for the sub-areas beneath the top position of the face, and height of the face is determined proportional to the face width.
2~g~
However this approach doesn t directly detect the outline of the face, and doesn t extract correctly the facial region for fine encoding. Further this article doesn t suygest encoding only the facial part to reduce unnecessary information caused by noise in the background part.
SUMMARY OF THE INVENTION
As with video telephones, while a speaker is in motion, the background part is still. According to the present invention, a moving interest part such as the face of a speaker is detected by using differential image between frames. Further th.~s detection is corrected based on the movement direction of the interest part. Responsive to the corrected detection, only a part of a motion picture signal corresponding to the interest part is exactly extracted and encoded. As a result, unnecessary encoding information generated by noise in the background can be obviated and the efficiency can be improved.
How to extract an interest part for example a speaker part is described with reference to the drawings. Suppose a speaXer moves as shown in FIG.2 from time "t0" to time "tl", and from time "tl" to time "t2". At first, regions having difference between the frames at the times tl and t2 are detected as illustrated in FIG.3 with hatching. Hatching portions in the background are caused by differential 2~)~81~rl3 signals generated from noise in the background. Ne~.t, the ~rame is divided into blocks of picture elements in m rows x n columns, and each block is determined as a valid if the sum of the absolute values o~ differential signals in the block is larger than or equal to a predetermined first threshold value, and the block is determined as invalid if the sum is less than the threshold value. A valid block map thus obtained is illustrated in FIG.4B, in which valid blocks are indicated with hatching.
Now suppose a valid block map obtained for frames at times tO and tl is illustrated in FIG.4A an~ the block is hereinafter referred to as a fifth valid block map or complete valid block map. And while the valid block map for the current frame, or the first valid block value is weighted With a first weight value, the valid block for the previous frame or the fifth valid block map is weighted with a second weight value. An example of the weighting is described as follows.
Suppose for example the weight value for valid blocks of the previous frame is "1" and that for its invalid blocks is "0", and the weight value for valid blocks of the current frame is "2" and that for its invalid blocks is "0" like that for the invalid blocks of the previous blocks. After weighting the valid block maps of the current and previous ~rames as mentioned above, the valid block maps are summed up to obtain a second valid block map. The second valid . .
:
: , , :
, 20~g~7~
block map is illustrated in EIG.5.
Next, segmentation is performed based on the composed valid block map. An example of the segmentation is described with reference to EIG's 5, 6, 7, and 8. In the valid block map, each block to be processed is set as valid if the sum of the values of the block and the eight neighbor blocks, that is, a block "k" and blocks "a", "b", "c", "d", "e", "f", "g", and "h" in FIG.7, is larger than or equal to a predetermined threshold value, and the block to be processed is set as invalid if the sum value is less than the threshold value.
A third valid block map is accordingly obtained as illustrated in FIG's.6A and 6B.
However, the above segmentation which uses the same threshold for all the blocks may make some blocks in the moving direction side area to b~ set as valid, changed to invalid and some portions of the person figure is dropped off as shown in FIG.6B, and an encoded:image~is degraded.
.
For avoiding that deterioration, values of the most left side valid blocks along horizontal block lines in the third valid block map are summed up, and likewise values of the most right side valid blocks along the horizontal block lines in the map are summed up. The difference between the left side valid block sum value and the right side valid block sum value is then calculated, and if the absolute 213~8~7~
value of the difference is larger than a predetermined third threshold value, then the valid blocks on the side having the larger sum value are weighted with a third weight value to avoid the valid area from dropping off.
As seen from the illustration of EIG.6A, the sum of the values of the most left side valid blocks along the horizontal block lines is "6" and the sum of the values of the most right side valid blocks along the horizontal block lines is "13". Suppose the absolute value of the dif~erence between the left and right side sums "7" (=13-6) is larger than the threshold. Then, the valid blocks on the right side are weighted with the third weight. In this case, the most right side valid blocks along the horizontal block lines are added by one and another valid block map is obtained as illustrated in FIG.8A.
After the weighting with the third weight as shown in FIG.
8A, each block in the third valid block map is set as valid with reference to the neighbor blocks if the sum value o~
the block and the neighbor blocks is larger than or equal to a predetermined fourth threshold value, and the block is set as invalid if the sum is less than the fourth threshold value. As a result, a fourth valid block map is obtained as - illustrated in FIG.8B in which valid areas are indicated with hatching.
If the absolute value of the difference between the sums of 2~G8~7~
the values of the blocks on the left and right sides is less than the third threshold value, there is no significant mGvement direction. Accordingly, the weighting with the third weight is not performed. And each block in the third valid block ~ap is set as valid wi-th reference to the neiyhbor blocks if the sum value of the block and the neighbor blocks is larger than or equal to a predetermined fifth threshold value, and the block is set as invalid if the sum is less than the fifth threshold value. As a result, the fourth valid block map is obtained.
As shown in FIG.8B, there can be isolated invalid blocks in the valid region. If ancoding is applied to only valid blocks in the fourth valid block map, the isolated invalid blocks in the valid region are not encoded, and an image of the previous frame remains at the isolated invalid blocks and causes discontinuation with the surrounding blocks.
This may result in significant deterioration of the encoded image. For resolving the issue, the same processing is applied to the isolated invalid blocks as used for the segmentation above mentioned. More particularly, in the fourth valid block map, each invalid block is checked with reference to the neighbor blocks, and if the sum of the values of the neighbor blocks is larger than or equal to a predetermined sixth threshold value, then the invalid block is filled with a value indicating validity. By this operation, the area of isolated invalid blocks in FIG.8B is removed, and a fifth valid block map is obtained. FIG.9 illustrate the fifth valid block map wherein the valid blocks are indicated with hatching.
Thereafter, the area of valid blocks or the speaker area of FIG.9 is encoded by using interframe and/or inframe correlations, and unnecessary information caused by the noise in the background is accordingly reliably removed. Therefore coding efflciency is improved.
The above threshold and weightlng values are determlned based on statlstical experiments. The arrangement of reference picture elements for the segmentatlon and the isolated block removal can be varled from the above.
In accordance with the present inventlon, there is provided a motlon plcture signal encoding method utllizing a correlation between images, comprising the following steps of:
generating a first valid block map for each frame, by dividing images of said frame and corresponding reference frame into blocks of a plurality of picture elements, detecting for each of said blocks difference between said images of said frame and said reference frame, making each of said blocks a valid block if said dlfference of sald block is larger than or equal to a predetermined first threshold value, and making said block an invalid block if said difference of said block is less than said first threshold value;
generating a second valid block map for each of said frames, by weighting said first valid block map with a first weighting factor, weighting a complete valld block map generated for a previous frame with a second weighting factor, and composing said weighted first and complete valid block 747~0-1 maps by sum operations;
generating a third valid block map for each of sald frames, by making each of said valid block of said second valid block map an invalid block if a sum of values of said valid block and neighbor blocks is less than a predetermined second threshold value;
generating a fourth valid block map for each of said frames, by, obtaining a sum of values of valid blocks at the most left locations along horizontal block lines in said third valid block map, obtaining a sum of values of valid blocks at the most right locations along horizontal block lines in said third valid block map, if a difference between said sums is larger than or equal to a third threshold value then weighting valid blocks at the most left or rlght locations having the greater sum with a third weighting factor and further maklng each block of sald thlrd valld block map thus weighted a valld block provided that a sum of values of sald block and neighbor blocks is larger than or equal to a fourth threshold value and making each of said block of said third valid block map thus weighted an invalid block provided that a sum of values of said block and neighbor blocks is less than said fourth threshold value to generate said fourth valid block map;
if said difference between said sums is less than said third threshold value then making each block of said third valid block map a valid block provided that a sum of values of -8a-747gO-l said block and neighbor blocks is larger than or equal to a fifth threshold value and making each of sald block of said third valid block map an invalid block provided that a sum of values of said block and nelghbor blocks is less than said fifth threshold value to generate said fourth valid block map;
generating said complete valid block map for each of sald frames, by making each of invalid blocks of said fourth valid block map a valid block if a sum of values of neighbor blocks is larger than or equal to a predetermined sixth threshold ~0 value; and encoding only a part of said image of said frame corresponding to valid blocks of said complete valid block map by using correlations between images and/or in a image.
In accordance with another aspect of the invention, there is provided a motion picture signal encoding apparatus utilizing a correlation between images, comprising:
means for generating a first valid block map for each frame, by dividing images of said frame and corresponding reference frame into blocks of a plurality of picture elements, detecting for each of said blocks difference between said images of said frame and said reference frame, making each of said blocks a valid block if said difference of said block is larger than or equal to a predetermined first threshold value, and making said block an invalid block if said difference of said block is less than said first threshold value;
means for generating a second valid block map for each of said frames, by weighting said first valid block map with a -8b-first weightinq factor, weighting a complete valid block map generated for a previous frame with a second factor, and composlng said weighted first and complete valid block maps by sum operations;
means for generating a third valid block map for each of said frames, by making each of said valld blocks of said second valid block map an invalid block if a sum of values of saidvalid block and neighbor blocks is less than a predetermined second threshold value;
means for generating a fourth valid block map for each of said frames, by, obtaining a sum of values of valid blocks at the most left locations along horizontal block lines in sald third valid block map, obtaining a sum of values of valid blocks at the most right locations along horizontal block lines in sald third valid block map, if a difference between said sums is larger than or equal to a third threshold value then welghting valid blocks at the most left or right locatlons havlng the greater sum with a third weighting factor and further making each block of said third valid block map thus weighted a valid block provided that a sum of values of said block and nelghbor blocks is larger than or equal to a fourth threshold value and making each of said blocks of said third valid block map thus weighted an invalid block provided that a sum of values of said block and neighbor blocks is less than said fourth threshold value to generate said fourth valid block map;
if said difference between said sums is less than said -8c-thlrd threshold value then making each block of said third valid block map a valid block provlded that a sum of values of sald block and neighbor blocks is larger than or equal to a fifth threshold value and making each of said blocks of said third valid block map an lnvalld block provlded that a sum of values of sald block and nelghbor blocks is less than said fifth threshold value to generate said fourth valid block map;
means for generatlng sald complete valld block map for each of said frames, by maklng each of invalid blocks of said fourth valld block map a valld block lf a sum of values of nelghbor blocks ls larger than or equal to a predetermined slxth threshold value; and means for encoding only a part of said lmage of sald frame correspondlng to valld blocks of sald complete valld block map by using correlatlons between lmages and/or ln an image.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 ls a block diagram illustrating an embodiment of the present inventlon.
FIG. 2 ls a drawing for lllustratlng the prlnclples of the present invention, in which a speaker is shown as moving from a tlme "t0" to "tl", from the time "tl" to "t2".
FIG. 3 is a drawlng for illustrating the principles of the present inventlon, whlch shows areas having a significant dlfference between the frames at the times "tl"
and "t2".
FIG.4 ls a drawlng for lllustratlng the prlnclples of the -8d-~g~
present invention, which shows a first valid block map (B) and a fifth valid block map of the previous frame (A).
FIG.5 is a drawing for illustrating the principles of the present invention, which shows a second valid block map.
FIG.6 is a drawing for illustrating the principles of the present invention, which shows a third valid block map (A), and an example incomplete encoding image caused by the third valid block map (B).
FIG.7 is a drawing for illustrating the principles of the present invention, which shows a block to be processed and neighbor blocks for segmentation.
FIG.8 is a drawing for illustrating the principles of the present invention, which shows a fourth valid block map (A), and the valid area in the fourth yalid block map (B).
FIG.9 is a drawing for illustrating the principles of the present invention, which shows the valid area in a fifth valid block map.
DETAILED DESCRIPTION OF A PREFERRED EMBODIMENT
One embodiment of the present invention is now described with reference to the drawings.
2~8~7~
FIG.1 shows the embodiment, in which an inputted motion picture signal is fed to a valid/invalid detection section 1 and a delay element 10 via a line 11. The valid/invalid detection section 1 stores a motion picture signal of a previous frame, and acquires the frame differential signal between the motion picture signal of the previous frame and that newly inputted via the line 11, and divides the frame differential signal into blocks of picture elements in m rows and n columns, and then calculates for each block the absolute value of the sum of frame differential values in the block. The valid/invalid detection section 1 further labels each block with "valid block" if the absolute value of the sum of the frame differential values in the block is larger than or equal to a predetermined first threshold value, and labels the block with "invalid block" if the absolute value of the sum of the differential values is less than the first threshold value, and accordingly obtains a first valid block map. The absolute value of the sum of the differential values can be replaced with a sum of squares, or the number of picture elements having values exceeding a threshold value.
The valid block signal obtained through the valid/invalid detection section 1 is fed to a weight section 2 which in turn per~orms a predetermined first weightlng operation on valid blocks in the valid block map fed from the valid/invalid detection section 1 to modify the values of the valid blocks. The first valid block map weighted within ,, ~ ,, , ~8~73 the weiyht section 2 is fed to a sum section 4 which sums the first valid block map thus fed and a fifth valid block map to generate a second valid block map thus weighted. The second valid block map obtained in the sum section 4 is fed to a segmentation section 5.
The segmentation section 5 calculates for every block in the second valid block map the sum of the values of the block and neighbor blocks, and makes the block an invalid block if its sum is less than a predetermined third threshold value to obtain a third valid block map.
The values in the third valid block map fed from the segmentation section 5 are checked along each horizontal block line in the map , from the left to the right, and the value of the first coming valid block is determined as the value at the left edge of the block line. Likewise, the values in the block map are checked along each horizontal block line, from the right to the left, and the value of the ! first coming valid block is determined as the value of the right edge of the block line. Thereafter, the sum of the values o~ the valid block on the left edges and that on the right edges are obtained, and the difference between the sums of the values of the valid blocks on the left edges and right edges is then calculated. If the absolute value of the difference is larger than or equal to a predetermined third threshold, then the valid blocks on the edge having a larger sum are weighted in a weight section 6 by addition of 2 ~ 7 3 a predetermined value.
The third valid block map which has undergone the third weighting operation is fed to a another segmentation section 7 in which each block is made a valid block if the sum of the values of the block and neighbor blocks is larger than or equal to a predetermined fourth threshold and the block is made an invalid block if the sum is less than the fourth threshold to generates a fourth valid block map.
When the difference between the sums of values of the valid blocks on the left and right sides is less than the third threshold value, in the segmentation section 7, each block is made a valid block if the sum of the values of the block and neighbor blocks is larger than or equal to a predetermined fifth threshold and the block is made an invalid block if the sum is less than the fifth threshold to generates the fourth valid block map.
As illustrated in FIG. 6, suppose that a block "k" is a block to be processed and that neighbor blocks are for example blocks "a", "b", "c", "d", "e", "~", "g", and "h".
If the sum of the their values is larger than or equal to the predetermined threshold value, then the block "k" is made a valid block, and the sum is less that the threshold value, then the block "k" is made an invalid block.
The fourth valid block map obtained through the segmentation ~g~7~
section 7 is supplied to an isolated invalid block removal section 8. The removal block section 8 removes isolated invalid blocks or invalid blocks in a valid block area to connect valid blocks. This process is performed with reference to neighbor blocks like the segmentation, by changing an invalid block to be processed into a valid block if the sum of values of the block and neighbor blocks is larger than or equal to a predetermined sixth threshold value. Thus, a fifth valid block map in which isolated invalid blocks have been removed is obtained.
The fifth valid block map thus obtained is supplied to another weight section 3 which performs a predetermined second weighting operation on valid blocks in the fifth valid block map to modify the values of the valid blocks, and supplies the modified fifth valid block map to the sum section 4 upon the next frame processing.
The delay section lO compensates an inputted motion picture signal for the time delay from the time when the motion picture signal is inputted to the time the fifth valid block map is supplied to an encoding section 9, and arraigns the timing of the fifth valid block map and the inputted signal.
The time compensated signal outputted from the delay section 10 is fed to the encoding section 9.
The encoding section 9 does encode only a part of the motion picture signal inputted from the delay section 10 ~, .
~6~7~
corresponding to the valid blocks of the fifth valid block map fed from the isolated invalid block removal section 8, that is, corresponding to a portion indicated as a speaker area, but doesn t encode the background part indicated with the invalid blocks. For the encoding, methods utilizing a correlation between frames, such as motion compensation, methods utilizing a correlation in a frame such as orthoganal conversion, and methods utiliziny correlations between frames and/or in a frame can be used.
The above threshold values are determined according to values statistically obtained beforehand.
As seen from the above detailed description of the preferred embodiment, according to the present invention, only a speaker area obtained by segmentation is encoded, and unnecessary information caused by noise in the background is removed, and the encoding efficiency is accordingly irnproved.
Although the embodiment of the present invention have been described, it is und0rstood that numerous variations may be made in accordance with the principles of the invention.
Claims (5)
1. A motion picture signal encoding method utilizing a correlation between images, comprising the following steps of:
generating a first valid block map for each frame, by dividing images of said frame and corresponding reference frame into blocks of a plurality of picture elements, detecting for each of said blocks difference between said images of said frame and said reference frame, making each of said blocks a valid block if said difference of said block is larger than or equal to a predetermined first threshold value, and making said block an invalid block if said difference of said block is less than said first threshold value;
generating a second valid block map for each of said frames, by weighting said first valid block map with a first weighting factor, weighting a complete valid block map generated for a previous frame with a second weighting factor, and composing said weighted first and complete valid block maps by sum operations;
generating a third valid block map for each of said frames, by making each of said valid block of said second valid block map an invalid block if a sum of values of said valid block and neighbor blocks is less than a predetermined second threshold value;
generating a fourth valid block map for each of said frames, by, obtaining a sum of values of valid blocks at the most left locations along horizontal block lines in said third valid block map, obtaining a sum of values of valid blocks at the most right locations along horizontal block lines in said third valid block map, if a difference between said sums is larger than or equal to a third threshold value then weighting valid blocks at the most left or right locations having the greater sum with a third weighting factor and further making each block of said third valid block map thus weighted a valid block provided that a sum of values of said block and neighbor blocks is larger than or equal to a fourth threshold value and making each of said block of said third valid block map thus weighted an invalid block provided that a sum of values of said block and neighbor blocks is less than said fourth threshold value to generate said fourth valid block map;
if said difference between said sums is less than said third threshold value then making each block of said third valid block map a valid block provided that a sum of values of said block and neighbor blocks is larger than or equal to a fifth threshold value and making each of said block of said third valid block map an invalid block provided that a sum of values of said block and neighbor blocks is less than said fifth threshold value to generate said fourth valid block map;
generating said complete valid block map for each of said frames, by making each of invalid blocks of said fourth valid block map a valid block if a sum of values of neighbor blocks is larger than or equal to a predetermined sixth threshold value; and encoding only a part of said image of said frame corresponding to valid blocks of said complete valid block map by using correlations between images and/or in a image.
generating a first valid block map for each frame, by dividing images of said frame and corresponding reference frame into blocks of a plurality of picture elements, detecting for each of said blocks difference between said images of said frame and said reference frame, making each of said blocks a valid block if said difference of said block is larger than or equal to a predetermined first threshold value, and making said block an invalid block if said difference of said block is less than said first threshold value;
generating a second valid block map for each of said frames, by weighting said first valid block map with a first weighting factor, weighting a complete valid block map generated for a previous frame with a second weighting factor, and composing said weighted first and complete valid block maps by sum operations;
generating a third valid block map for each of said frames, by making each of said valid block of said second valid block map an invalid block if a sum of values of said valid block and neighbor blocks is less than a predetermined second threshold value;
generating a fourth valid block map for each of said frames, by, obtaining a sum of values of valid blocks at the most left locations along horizontal block lines in said third valid block map, obtaining a sum of values of valid blocks at the most right locations along horizontal block lines in said third valid block map, if a difference between said sums is larger than or equal to a third threshold value then weighting valid blocks at the most left or right locations having the greater sum with a third weighting factor and further making each block of said third valid block map thus weighted a valid block provided that a sum of values of said block and neighbor blocks is larger than or equal to a fourth threshold value and making each of said block of said third valid block map thus weighted an invalid block provided that a sum of values of said block and neighbor blocks is less than said fourth threshold value to generate said fourth valid block map;
if said difference between said sums is less than said third threshold value then making each block of said third valid block map a valid block provided that a sum of values of said block and neighbor blocks is larger than or equal to a fifth threshold value and making each of said block of said third valid block map an invalid block provided that a sum of values of said block and neighbor blocks is less than said fifth threshold value to generate said fourth valid block map;
generating said complete valid block map for each of said frames, by making each of invalid blocks of said fourth valid block map a valid block if a sum of values of neighbor blocks is larger than or equal to a predetermined sixth threshold value; and encoding only a part of said image of said frame corresponding to valid blocks of said complete valid block map by using correlations between images and/or in a image.
2. A motion picture signal encoding method according to the CLAIM.1, wherein said threshold values are determined statistically.
3. A motion picture signal encoding method according to the CLAIM.2 wherein said neighbor blocks are eight blocks adjacent to a block to be processed.
4. A motion picture signal encoding method comprising the following steps of:
receiving an inputted motion picture signal for each frame;
locating an area covering an interest part from said motion picture signal for each of said frame;
responsive to said located area, using the method of claim 1 for extracting said interest part from said motion picture signal for each of said frames and for encoding only said interest part of said motion picture signal thus extracted by using correlations between images and/or in an image;
wherein said step of locating said area comprises detecting a movement direction of said interest part; and modifying said located area in the fourth effective block map based on said movement direction thus detected.
receiving an inputted motion picture signal for each frame;
locating an area covering an interest part from said motion picture signal for each of said frame;
responsive to said located area, using the method of claim 1 for extracting said interest part from said motion picture signal for each of said frames and for encoding only said interest part of said motion picture signal thus extracted by using correlations between images and/or in an image;
wherein said step of locating said area comprises detecting a movement direction of said interest part; and modifying said located area in the fourth effective block map based on said movement direction thus detected.
5. A motion picture signal encoding apparatus utilizing a correlation between images, comprising:
means for generating a first valid block map for each frame, by dividing images of said frame and corresponding reference frame into blocks of a plurality of picture elements, detecting for each of said blocks difference between said images of said frame and said reference frame, making each of said blocks a valid block if said difference of said block is larger than or equal to a predetermined first threshold value, and making said block an invalid block if said difference of said block is less than said first threshold value;
means for generating a second valid block map for each of said frames, by weighting said first valid block map with a first weighting factor, weighting a complete valid block map generated for a previous frame with a second factor, and composing said weighted first and complete valid block maps by sum operations;
means for generating a third valid block map for each of said frames, by making each of said valid block of said second valid block map an invalid block if a sum of values of said valid block and neighbor blocks is less than a predetermined second threshold value;
means for generating a fourth valid block map for each of said frames, by, obtaining a sum of values of valid blocks at the most left locations along horizontal block lines in said third valid block map, obtaining a sum of values of valid blocks at the most right locations along horizontal block lines in said third valid block map, if a difference between said sums is larger than or equal to a third threshold value then weighting valid blocks at the most left or right locations having the greater sum with a third weighting factor and further making each block of said third valid block map thus weighted a valid block provided that a sum of values of said block and neighbor blocks is larger than or equal to a fourth threshold value and making each of said block of said third valid block map thus weighted an invalid block provided that a sum of values of said block and neighbor blocks is less than said fourth threshold value to generate said fourth valid block map;
if said difference between said sums is less than said third threshold value then making each block of said third valid block map a valid block provided that a sum of values of said block and neighbor blocks is larger than or equal to a fifth threshold value and making each of said block of said third valid block map an invalid block provided that a sum of values of said block and neighbor blocks is less than said fifth threshold value to generate said fourth valid block map, means for generating said complete valid block map for each of said frames, by making each of invalid blocks of said fourth valid block map a valid block if a sum of values of neighbor blocks is larger than or equal to a predetermined sixth threshold value; and means for encoding only a part of said image of said frame corresponding to valid blocks of said complete valid block map by using correlations between images and/or in a image.
means for generating a first valid block map for each frame, by dividing images of said frame and corresponding reference frame into blocks of a plurality of picture elements, detecting for each of said blocks difference between said images of said frame and said reference frame, making each of said blocks a valid block if said difference of said block is larger than or equal to a predetermined first threshold value, and making said block an invalid block if said difference of said block is less than said first threshold value;
means for generating a second valid block map for each of said frames, by weighting said first valid block map with a first weighting factor, weighting a complete valid block map generated for a previous frame with a second factor, and composing said weighted first and complete valid block maps by sum operations;
means for generating a third valid block map for each of said frames, by making each of said valid block of said second valid block map an invalid block if a sum of values of said valid block and neighbor blocks is less than a predetermined second threshold value;
means for generating a fourth valid block map for each of said frames, by, obtaining a sum of values of valid blocks at the most left locations along horizontal block lines in said third valid block map, obtaining a sum of values of valid blocks at the most right locations along horizontal block lines in said third valid block map, if a difference between said sums is larger than or equal to a third threshold value then weighting valid blocks at the most left or right locations having the greater sum with a third weighting factor and further making each block of said third valid block map thus weighted a valid block provided that a sum of values of said block and neighbor blocks is larger than or equal to a fourth threshold value and making each of said block of said third valid block map thus weighted an invalid block provided that a sum of values of said block and neighbor blocks is less than said fourth threshold value to generate said fourth valid block map;
if said difference between said sums is less than said third threshold value then making each block of said third valid block map a valid block provided that a sum of values of said block and neighbor blocks is larger than or equal to a fifth threshold value and making each of said block of said third valid block map an invalid block provided that a sum of values of said block and neighbor blocks is less than said fifth threshold value to generate said fourth valid block map, means for generating said complete valid block map for each of said frames, by making each of invalid blocks of said fourth valid block map a valid block if a sum of values of neighbor blocks is larger than or equal to a predetermined sixth threshold value; and means for encoding only a part of said image of said frame corresponding to valid blocks of said complete valid block map by using correlations between images and/or in a image.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP3133540A JPH04334188A (en) | 1991-05-08 | 1991-05-08 | Coding system for moving picture signal |
| JP3-133540 | 1991-05-08 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CA2068073A1 CA2068073A1 (en) | 1992-11-09 |
| CA2068073C true CA2068073C (en) | 1998-10-06 |
Family
ID=15107203
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CA 2068073 Expired - Fee Related CA2068073C (en) | 1991-05-08 | 1992-05-06 | Encoding of motion picture signal |
Country Status (5)
| Country | Link |
|---|---|
| US (1) | US5258836A (en) |
| EP (1) | EP0514053B1 (en) |
| JP (1) | JPH04334188A (en) |
| CA (1) | CA2068073C (en) |
| DE (1) | DE69215577T2 (en) |
Families Citing this family (57)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3093499B2 (en) * | 1992-03-06 | 2000-10-03 | 三菱電機株式会社 | Scene change detection device |
| JP2814832B2 (en) * | 1992-04-08 | 1998-10-27 | 日本電気株式会社 | Image encoding / decoding device |
| FR2690031A1 (en) * | 1992-04-14 | 1993-10-15 | Philips Electronique Lab | Segmentation device of images. |
| JP3526057B2 (en) * | 1992-07-31 | 2004-05-10 | キヤノン株式会社 | Video transmission equipment |
| JPH06153180A (en) * | 1992-09-16 | 1994-05-31 | Fujitsu Ltd | Picture data coding method and device |
| JPH06209466A (en) * | 1992-10-07 | 1994-07-26 | Canon Inc | Motion vector detector |
| KR0128857B1 (en) * | 1992-12-23 | 1998-04-10 | 배순훈 | Window signal generation device |
| CA2119327A1 (en) * | 1993-07-19 | 1995-01-20 | David Crawford Gibbon | Method and means for detecting people in image sequences |
| US5610657A (en) * | 1993-09-14 | 1997-03-11 | Envistech Inc. | Video compression using an iterative error data coding method |
| JP2591441B2 (en) * | 1993-09-28 | 1997-03-19 | 日本電気株式会社 | Moving vector detection device for moving image signal |
| GB9325073D0 (en) * | 1993-12-07 | 1994-02-02 | Eidos Plc | Improvements in or relating to video processing systems |
| US6052414A (en) * | 1994-03-30 | 2000-04-18 | Samsung Electronics, Co. Ltd. | Moving picture coding method and apparatus for low bit rate systems using dynamic motion estimation |
| US5519436A (en) * | 1994-06-21 | 1996-05-21 | Intel Corporation | Static image background reference for video teleconferencing applications |
| EP1098527A1 (en) * | 1994-11-04 | 2001-05-09 | Matsushita Electric Industrial Co., Ltd. | Picture coding apparatus and decoding apparatus |
| KR100186979B1 (en) * | 1994-12-28 | 1999-05-01 | 김주용 | Image data encoding method for uncovered background of object |
| JP3086396B2 (en) * | 1995-03-10 | 2000-09-11 | シャープ株式会社 | Image encoding device and image decoding device |
| KR0171120B1 (en) * | 1995-04-29 | 1999-03-20 | 배순훈 | Method and apparatus for determining motion region in video coding technique using feature point based motion compensation |
| US5959673A (en) * | 1995-10-05 | 1999-09-28 | Microsoft Corporation | Transform coding of dense motion vector fields for frame and object based video coding applications |
| US5781198A (en) * | 1995-12-22 | 1998-07-14 | Intel Corporation | Method and apparatus for replacing a background portion of an image |
| US5799113A (en) * | 1996-01-19 | 1998-08-25 | Microsoft Corporation | Method for expanding contracted video images |
| US5787203A (en) * | 1996-01-19 | 1998-07-28 | Microsoft Corporation | Method and system for filtering compressed video images |
| US5778098A (en) * | 1996-03-22 | 1998-07-07 | Microsoft Corporation | Sprite coding |
| US5982438A (en) * | 1996-03-22 | 1999-11-09 | Microsoft Corporation | Overlapped motion compensation for object coding |
| CN1165180C (en) | 1996-05-17 | 2004-09-01 | 松下电器产业株式会社 | Image coding method |
| AU4131897A (en) * | 1996-09-12 | 1998-04-02 | British Technology Group Plc | Object-oriented video system |
| US6075875A (en) * | 1996-09-30 | 2000-06-13 | Microsoft Corporation | Segmentation of image features using hierarchical analysis of multi-valued image data and weighted averaging of segmentation results |
| US5748789A (en) * | 1996-10-31 | 1998-05-05 | Microsoft Corporation | Transparent block skipping in object-based video coding systems |
| US6370192B1 (en) * | 1997-11-20 | 2002-04-09 | Hitachi America, Ltd. | Methods and apparatus for decoding different portions of a video image at different resolutions |
| US6400831B2 (en) | 1998-04-02 | 2002-06-04 | Microsoft Corporation | Semantic video object segmentation and tracking |
| US6711278B1 (en) * | 1998-09-10 | 2004-03-23 | Microsoft Corporation | Tracking semantic objects in vector image sequences |
| US6983018B1 (en) | 1998-11-30 | 2006-01-03 | Microsoft Corporation | Efficient motion vector coding for video compression |
| US6563953B2 (en) | 1998-11-30 | 2003-05-13 | Microsoft Corporation | Predictive image compression using a single variable length code for both the luminance and chrominance blocks for each macroblock |
| US7200275B2 (en) | 2001-12-17 | 2007-04-03 | Microsoft Corporation | Skip macroblock coding |
| US7085420B2 (en) * | 2002-06-28 | 2006-08-01 | Microsoft Corporation | Text detection in continuous tone image segments |
| US7072512B2 (en) * | 2002-07-23 | 2006-07-04 | Microsoft Corporation | Segmentation of digital video and images into continuous tone and palettized regions |
| US7426308B2 (en) | 2003-07-18 | 2008-09-16 | Microsoft Corporation | Intraframe and interframe interlace coding and decoding |
| US7499495B2 (en) | 2003-07-18 | 2009-03-03 | Microsoft Corporation | Extended range motion vectors |
| US7738554B2 (en) | 2003-07-18 | 2010-06-15 | Microsoft Corporation | DC coefficient signaling at small quantization step sizes |
| US10554985B2 (en) | 2003-07-18 | 2020-02-04 | Microsoft Technology Licensing, Llc | DC coefficient signaling at small quantization step sizes |
| US7609763B2 (en) | 2003-07-18 | 2009-10-27 | Microsoft Corporation | Advanced bi-directional predictive coding of video frames |
| US20050013498A1 (en) | 2003-07-18 | 2005-01-20 | Microsoft Corporation | Coding of motion vector information |
| US7577200B2 (en) | 2003-09-07 | 2009-08-18 | Microsoft Corporation | Extended range variable length coding/decoding of differential motion vector information |
| US7567617B2 (en) | 2003-09-07 | 2009-07-28 | Microsoft Corporation | Predicting motion vectors for fields of forward-predicted interlaced video frames |
| US7092576B2 (en) * | 2003-09-07 | 2006-08-15 | Microsoft Corporation | Bitplane coding for macroblock field/frame coding type information |
| US7616692B2 (en) | 2003-09-07 | 2009-11-10 | Microsoft Corporation | Hybrid motion vector prediction for interlaced forward-predicted fields |
| US7317839B2 (en) | 2003-09-07 | 2008-01-08 | Microsoft Corporation | Chroma motion vector derivation for interlaced forward-predicted fields |
| US7724827B2 (en) | 2003-09-07 | 2010-05-25 | Microsoft Corporation | Multi-layer run level encoding and decoding |
| US7620106B2 (en) | 2003-09-07 | 2009-11-17 | Microsoft Corporation | Joint coding and decoding of a reference field selection and differential motion vector information |
| US7606308B2 (en) | 2003-09-07 | 2009-10-20 | Microsoft Corporation | Signaling macroblock mode information for macroblocks of interlaced forward-predicted fields |
| US7599438B2 (en) | 2003-09-07 | 2009-10-06 | Microsoft Corporation | Motion vector block pattern coding and decoding |
| US8064520B2 (en) | 2003-09-07 | 2011-11-22 | Microsoft Corporation | Advanced bi-directional predictive coding of interlaced video |
| US7623574B2 (en) | 2003-09-07 | 2009-11-24 | Microsoft Corporation | Selecting between dominant and non-dominant motion vector predictor polarities |
| JP5008269B2 (en) * | 2005-04-08 | 2012-08-22 | キヤノン株式会社 | Information processing apparatus and information processing method |
| US9077960B2 (en) | 2005-08-12 | 2015-07-07 | Microsoft Corporation | Non-zero coefficient block pattern coding |
| US8254455B2 (en) | 2007-06-30 | 2012-08-28 | Microsoft Corporation | Computing collocated macroblock information for direct mode macroblocks |
| US7925774B2 (en) | 2008-05-30 | 2011-04-12 | Microsoft Corporation | Media streaming using an index file |
| US8189666B2 (en) | 2009-02-02 | 2012-05-29 | Microsoft Corporation | Local picture identifier and computation of co-located information |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4383272A (en) * | 1981-04-13 | 1983-05-10 | Bell Telephone Laboratories, Incorporated | Video signal interpolation using motion estimation |
| JPS6075184A (en) * | 1983-09-30 | 1985-04-27 | Nec Corp | System and apparatus for coding dynamic picture signal |
| CA1287161C (en) * | 1984-09-17 | 1991-07-30 | Akihiro Furukawa | Apparatus for discriminating a moving region and a stationary region in a video signal |
| GB8710737D0 (en) * | 1987-05-06 | 1987-06-10 | British Telecomm | Video image encoding |
| EP0330455A3 (en) * | 1988-02-22 | 1990-07-04 | Kabushiki Kaisha Toshiba | Image encoding apparatus |
| JPH0220185A (en) * | 1988-07-08 | 1990-01-23 | Ricoh Co Ltd | Moving image transmission system |
| US4991009A (en) * | 1988-07-08 | 1991-02-05 | Ricoh Company, Ltd. | Dynamic image transmission system |
| DE3927516A1 (en) * | 1989-08-21 | 1991-02-28 | Ant Nachrichtentech | METHOD FOR FORMING A PREDICTION IMAGE |
| EP0476603B1 (en) * | 1990-09-20 | 1997-06-18 | Nec Corporation | Method and apparatus for coding moving image signal |
-
1991
- 1991-05-08 JP JP3133540A patent/JPH04334188A/en active Pending
-
1992
- 1992-05-01 EP EP19920303977 patent/EP0514053B1/en not_active Expired - Lifetime
- 1992-05-01 DE DE1992615577 patent/DE69215577T2/en not_active Expired - Fee Related
- 1992-05-05 US US07/878,471 patent/US5258836A/en not_active Expired - Fee Related
- 1992-05-06 CA CA 2068073 patent/CA2068073C/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| EP0514053A2 (en) | 1992-11-19 |
| JPH04334188A (en) | 1992-11-20 |
| CA2068073A1 (en) | 1992-11-09 |
| US5258836A (en) | 1993-11-02 |
| DE69215577T2 (en) | 1997-06-12 |
| EP0514053B1 (en) | 1996-12-04 |
| EP0514053A3 (en) | 1993-06-09 |
| DE69215577D1 (en) | 1997-01-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CA2068073C (en) | Encoding of motion picture signal | |
| EP0634873B1 (en) | Method to determine the motion vectors in small picture segments of a television picture | |
| US5751378A (en) | Scene change detector for digital video | |
| US5787199A (en) | Apparatus for detecting a foreground region for use in a low bit-rate image signal encoder | |
| US5572258A (en) | Motion compensation estimating device and method achieving improved motion compensation | |
| CA2051891A1 (en) | Method and apparatus for coding moving image signal | |
| US7460172B2 (en) | Frame interpolating method and apparatus thereof at frame rate conversion | |
| EP0528293B1 (en) | Apparatus for reducing quantization artifacts in an interframe hybrid coding system with motion compensation | |
| KR100488422B1 (en) | Grayscale-shaped information encoding / decoding device and method | |
| JPS6152084A (en) | Highly efficient encoder | |
| US5805221A (en) | Video signal coding system employing segmentation technique | |
| US5650828A (en) | Method and apparatus for detecting and thinning a contour image of objects | |
| US5696551A (en) | Three-dimension picture coding system | |
| US5978510A (en) | Method and apparatus for predictively coding contour of video signal | |
| JP2591441B2 (en) | Moving vector detection device for moving image signal | |
| Mizuki et al. | A binary block matching architecture with reduced power consumption and silicon area requirement | |
| JP3056120B2 (en) | Video signal shape information predictive coding method | |
| US5953079A (en) | Machine method for compensating for non-linear picture transformations, E.G. zoom and pan, in a video image motion compensation system | |
| GB2320989A (en) | Padding image blocks using substitute pixels for shape adaptive discrete cosine transform | |
| JPH089379A (en) | Motion vector detection method | |
| JP2586715B2 (en) | Video signal coding method | |
| JPH09182077A (en) | Image encoding method and device | |
| JP2586714B2 (en) | Video signal coding method | |
| KR0170866B1 (en) | Estimation device for moving region | |
| Panis et al. | The use of stereo and motion in a generic object-based coder |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| EEER | Examination request | ||
| MKLA | Lapsed |