IPC분류정보
국가/구분 |
United States(US) Patent
등록
|
국제특허분류(IPC7판) |
|
출원번호 |
US-0336984
(1982-01-04)
|
발명자
/ 주소 |
|
출원인 / 주소 |
|
대리인 / 주소 |
Fliesler, Dubb, Meyer & Lovejoy
|
인용정보 |
피인용 횟수 :
87 인용 특허 :
8 |
초록
▼
Video type information signals are compressed for transmission and reproduction by comparing corresponding blocks of time domain information signals from successive fields, converting a block of the time domain information signals to a transform domain signal represented by discrete cosine transform
Video type information signals are compressed for transmission and reproduction by comparing corresponding blocks of time domain information signals from successive fields, converting a block of the time domain information signals to a transform domain signal represented by discrete cosine transform coefficients when the difference between the corresponding blocks exceeds a block difference threshold, and encoding the transform domain coefficients for transmission to a decoding site. Corresponding blocks of time domain information signals from successive fields are compared by storing the successive fields in memory on a pixel by pixel basis, retrieving each block on a pixel by pixel basis, forming the difference between corresponding pixels from the successive blocks, squaring the resulting difference signal, summing the squares and dividing by the number of pixels per block. Successive fields are merged by weighted summing of corresponding pixels.
대표청구항
▼
1. A method for processing time domain information signals having a successive field format to effect substantial compression of said signals, said method comprising the steps of: comparing corresponding blocks of time domain information signals from successive fields to form a block difference s
1. A method for processing time domain information signals having a successive field format to effect substantial compression of said signals, said method comprising the steps of: comparing corresponding blocks of time domain information signals from successive fields to form a block difference signal; converting a block of said time domain information signals to transform domain signals including a D.C. coefficient representing the average intensity of a converted block and a plurality of discrete transform coefficients when said block difference signal exceeds a first variable parametric value; varying said first variable parametric value in accordance with the number of blocks selected for conversion; and encoding said transform domain coefficients for subsequent utilization. 2. The method of claim 1 wherein said step of comparing includes the steps of storing successive fields of said time domain information signals in a first memory device on a pixel by pixel basis, and retrieving said corresponding blocks from said memory device on a pixel by pixel basis. 3. The method of claim 2 further including the step of replacing a block pf pixel elements in said memory device with the corresponding block in a successive field when said difference exceeds said first variable parametric value. 4. The method of claim 2 wherein said step of storing includes the step of merging successive fields on a pixel by pixel basis. 5. The method of claim 4 wherein said step of merging is performed by summing corresponding pixels from successive fields in accordance with a predetermined weighting factor. 6. The method of claim 5 wherein the pixels from the first appearing merged field are weighted by a factor of 3/4 and the pixels from the later appearing field are weighted by a factor of 1/4. 7. The method of claim 1 wherein said step of converting is performed by transforming said block along a first direction and subsequently transforming said block along the orthogonal direction. 8. The method of claim 7 wherein said first direction corresponds to a horizontal line of raster scan information and said orthogonal direction corresponds to a vertical column of raster scan information. 9. The method of claim 7 wherein said step of converting includes the step of storing said transform coefficients in diagonal format in a second memory device. 10. The method of claim 9 further including the step of storing the field address of each transformed block in said second memory device. 11. A method for processing time domain information signals having a successive field format to effect substantial compression of said signals, said method comprising the steps of: comparing corresponding blocks of time domain information signals from successive fields to form a block difference signal, said step of comparing including the steps of storing successive fields of said time domain information signals in a first memory device on a pixel by pixel basis, retrieving said corresponding blocks from said memory device on a pixel by pixel basis, forming the difference between corresponding pixels from said successive blocks, squaring the resulting pixel difference signals, summing the squares of said resulting pixel difference signals, and dividing the resulting sum by a number of pixels per block to form said block difference signal; converting a block of said time domain information signals to transform domain signals including a D.C. coefficient representing the average intensity of a converted block and a plurality of discrete transform coefficients when said block difference signal exceeds a first variable parametric value; and encoding said transform domain coefficients for subsequent utilization. 12. The method of claim 11 wherein each block comprises a total of 64 pixels grouped in an 8 by 8 array. 13. A method for processing time domain information signals having a successive field format to effect substantial compression of said signals, said method comprising the steps of: comparing corresponding blocks of time domain information signals from successive fields to form a block difference signal; converting a block of said time domain information signals to transform domain signals including a D.C. coefficient representing the average intensity of a converted block and a plurality of discrete transform coefficients when said block difference signal exceeds a first variable parametric value; and encoding said transform domain coefficients for subsequent utilization, including the steps of providing a plurality of different code tables, dividing each of said plurality of discrete transform coefficients by a second variable parametric value to obtain a corresponding quantized coefficient, calculating the predictive value of each quantized coefficient, selecting one of said plurality of code tables in accordance with the predictive value, and generating a code representing said quantized coefficient from said selected table. 14. The method of claim 13 wherein said step of calculating is performed in accordance with the formula PM K =1/4C K +3/4PM K-1, where PM K is the predictive mean value of the K th quantized coefficient, C K is the value of the K th quantized coefficient and PM K-1 is the predictive mean value of the K-1 th quantized coefficient. 15. The method of claim 13 wherein said step of encoding includes the steps of providing a dedicated code table for encoding each DC coefficient and selecting said dedicated table for each said DC coefficient. 16. The method of claim 13 wherein said time domain information signals are color video signals having quadrature components, and wherein said step of encoding includes the steps of providing individual code tables for said quadrature components, calculating the average value of each quadrature component for the converted block, and selecting a code value representing said average value from the corresponding individual quadrature component table. 17. The method of claim 13 wherein said step of encoding further includes the step of generating an address code specifying the field address of the converted block. 18. The method of claim 17 wherein said step of generating includes the steps of providing an address code table having code values arranged in accordance with the following formula: If Δ k =1, code 1 bit (Δ k ) If Δ k <32, code 1 bit zero+5 bits (Δ k ) If Δ k ≥32, code 6 bits zero+10 bits (Δ k ) where Δ k =A k -A k-1 A k =numerical address of current block A k-1 =numerical address of most recently encoded block, and selecting said address code table for each said address code. 19. The method of claim 13 wherein said step of encoding further includes the steps of comparing said predictive value with a preselected fixed threshold value, and generating a run length code specifying the total number of successive quantized coefficients of zero value when said predictive value lies below said preselected fixed threshold value. 20. The method of claim 19 wherein said step of encoding further includes the step of generating an end of block code when the predictive values for successive remaining quantized coefficients in the converted block lie below said preselected fixed threshold value. 21. The method of claim 19 wherein said codes are multi-bit binary codes, and wherein said method further includes the steps of transferring said codes to a buffer in the order of generation, monitoring the number of bits transferred to said buffer, and varying said first and second parametric values in accordance with said number of transferred bits in order to minimize buffer overflow. 22. The method of claim 21 wherein said second parametric value is varied in accordance with the formula: ##EQU##D=D' K +K D ·BFN(B K -N/2) where BFN(X)=(X/N-|X|) D K =Distortion parameter for block K D' K =Filtered distortion parameter ##EQU##D' K =T·D' K-1 +(1-T)D K-1 where T=a constant (close to 1) K D =a constant B K =# of bits in buffer for block K N=Max. number of bits. 23. The method of claim 22 wherein said first parametric value is varied in accordance with the alternate formula: ##EQU##T K =T INIT for B LOW≤B K ≤B HIGH ##EQU##T K =T INIT +K R ·BFN(B LOW -B K ) for B K <B LOW ##EQU##T K =T INIT +K R ·BFN(B K -B HIGH ) for B K >B HIGH where T K =replenishment threshold for block K T INIT =initial threshold (about 5 for 8-bit input data) K R =multiplier constant (about 25-75) B LOW =low cutoff (about 0.1 of buffer) B HIGH =high cutoff (about 0.75 of buffer). 24. The method of claim 13 wherein each of said code tables comprises a Huffman code table. 25. A method of encoding transform coefficients representing time domain information signals having a successive field format prior to transmission over a communication link in order to effect substantial compression of said signals, said transform coefficients being arranged in a plurality of groups, each group representing an N by N block of field information signals, said method comprising the steps of: (a) providing a plurality of code tables: (b) generating a block address code from a first dedicated one of said code tables, said block address code representing the field address of the block represented by a group of said transform coefficients; (c) generating a DC coefficient code representing the average intensity of said block from a second dedicated one of said plurality of code tables; and (d) generating a succession of codes representing the remaining transform coefficients corresponding to said block by calculating the predictive value of each said remaining transform coefficient, selecting one of said plurality of code tables in accordance with said predictive value, and generating a code representing the corresponding transform coefficient from said selected table. 26. The method of claim 25 wherein said step of calculating is performed in accordance with the formula PM K =1/4 C K +3/4 PM K-1, where PM K is the predictive mean value of the K th coefficient, C K is the actual value of the K th coefficient and PM K-1 is the predictive mean value of the K-1 th coefficient. 27. The method of claim 25 wherein said time domain information signals are color video signals having quadrature components, and wherein said method of encoding includes the steps of providing individual code tables for said quadrature components, calculating the average value of each quadrature component for said block, and selecting a code value representing said average value from the corresponding individual quadrature component table. 28. The method of claim 25 wherein said method of encoding further includes the steps of comparing each said predictive value with a preselected threshold value, and generating a zero run length code specifying the total number of successive predictive values lying below said preselected threshold value. 29. The method of claim 28 wherein said codes are multi-bit binary codes, wherein said method of encoding further includes dividing each of said remaining transform coefficients by a variable parametric value to obtain corresponding quantized remaining coefficients for use in forming the predictive values, and wherein said method further includes the steps of transferring said codes to a buffer in the order of generation, monitoring the number of bits transferred to said buffer, and varying said variable parametric value in accordance with the following formula: ##EQU##D=D' K +K D ·BFN(B K -N/2) where BFN(X)=X/(N-|X|) D K =Distortion parameter for block K D' K =Filtered distortion parameter ##EQU##D' K =T·D' K-1 +(1-T)D K-1 where T=a constant (close to 1) K D =a constant B K =# of bits in buffer for block K N=Max. number of bits. 30. The method of claim 28 further including the step of generating an end of block code when the predictive values for successive remaining transform coefficients in said block lie below said preselected threshold value. 31. The method of claim 25 wherein said transform coefficients comprise discrete consine transform coefficients. 32. A method for processing time domain information signals for transmission over a communication link, said time domain information signals having a successive field format, said method comprising the steps of: comparing corresponding sub-field blocks of time domain information signals from successive fields to form a block difference signal; converting a sub-field block of said time domain information signals to a transform domain signal represented by a D.C. coefficient representing the average intensity of the converted block and a pluality of discrete transform coefficients when said block difference signal between the corresponding blocks exceeds a first variable parametric value; and generating a frame sync code signal indicating the beginning of a frame; generating a first control code signal B K representative of the fullness of a transmission rate buffer at the beginning of said frame; generating a second control code signal D K representative of a second variable parametric value at the beginning of said frame; and generating a plurality of block replenishment code symbols each representative of the value of transform coefficients corresponding to individual sub-field blocks having interfield block differences greater than said first variable parametric value, each said block replenishment code symbol including a block address code specifying the field address of the corresponding block, a D.C. code term representative of said D.C. coefficient of the corresponding block, and a plurality of coefficient code terms representative of the value of said plurality of discrete transform coefficients as quantified according to said second control code signal D K for the corresponding block. 33. The method of claim 32 wherein said time domain information signals are color video signals having quadrature components, and wherein said step of generating block replenishment code symbols includes the step of providing first and second color code terms in each of said plurality of said block replenishment code symbols representing the average value of each quadrature component for said corresponding block between said block address code and said DC code term. 34. The method of claim 32 wherein said step of generating block replenishment code symbols includes the step of providing a run length code term specifying the total number of successive transform coefficient zero values having a predictive mean value less than a preselected fixed threshold value. 35. The method of claim 32 wherein said step of generating block replenishment code symbols includes a step of providing an end of block code term when the predictive values for successive remaining transform coefficients in said corresponding block lie below a preselected fixed threshold value.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.