Image analysis – Image compression or coding – Quantization
Reexamination Certificate
1997-08-12
2003-09-16
Wu, Jingge (Department: 2623)
Image analysis
Image compression or coding
Quantization
Reexamination Certificate
active
06621934
ABSTRACT:
FIELD OF THE INVENTION
This invention concerns a digital video processor. Specifically, the invention concerns image compression method and apparatus for reducing the memory requirements of a digital video processor.
BACKGROUND OF THE INVENTION
The efficient use of memory is important in the design and operation of image processors. For example, consumer products such as television systems may use image processors including MPEG-2 signal processing. The MPEG (Motion Picture Experts Group) signal compression standard (ISO/IEC 13181-2, May 10, 1994) is a widely accepted image processing standard which is particularly attractive for use with satellite, cable and terrestrial broadcast systems employing high definition television (HDTV) processing among other forms of image processing. Products using high definition displays may require 96 Mbits or more of memory to temporarily store MPEG decoded frames prior to display. An MPEG processor requires these frames for motion estimation and compensation to reconstruct accurate images for display.
Systems which reconstruct images from MPEG decoded picture elements (pixels or pels) typically employ Differential Pulse Coded Modulation (DPCM). For DPCM, a processor generates a prediction value which anticipates the next pixel value. A summation network subtracts the prediction from the actual pixel value resulting in a difference which is used to represent the current pixel value as well as to anticipate the next pixel value. This difference, known as prediction error, is generally smaller than the original pixel or prediction values, so processing the difference rather than the original pixel value reduces system bandwidth requirements. The prediction error may have a positive or negative value. Ang et al., “Video Compression Makes Big Gains,” IEEE Spectrum, October 1991, generally describes an MPEG encoder and decoder employing DPCM processing.
Memory efficient image processors use less memory to store image frames by recoding (recompressing) the block data prior to storage. In the spacial domain, reducing the number of bits per pixel used to store the image frames adversely affects the picture quality if the pixels can not be accurately reconstructed to their original bit size. Artifacts may occur, especially in smooth areas of the image. Memory reduction image processors should accurately quantize and dequantize the MPEG decoded signal as efficiently and economically as possible. It is known to take advantage of human optical physiology by processing luminance and chrominance data differently. Optimizing compression laws for each type of data to account for the energy in the data as well as what the human eye can see is described in U.S. Pat. No. 4,575,749, Acampora et al. The Acampora patent discloses amplitude compression to reduce noise in television signals prior to transmission.
The present inventors recognize the desirability of providing an efficient data reduction system employing minimal hardware and software which will save memory and reduce the physical size of the processor while minimizing artifacts introduced into the reconstructed data. A system according to the present invention satisfies these objectives.
SUMMARY OF THE INVENTION
In accordance with the principles of the present invention, an image processor produces DPCM prediction error values to be quantized. Prior to quantization, the processor determines if the prediction error is positive or negative. If the prediction error value is positive, the value passes unchanged to a quantizer. If the prediction error value is negative, a bias value is added to the prediction error value to produce a positive number within the scope of the quantizer. Biased and unbiased prediction error values pass to the quantizer. Because all values received by the quantizer are positive and within the current quantizer limits, the quantization table used by the quantizer need not include quantization values for negative prediction error values. This reduces the scope of prediction error values by a factor of two and doubles quantization resolution.
In an exemplary embodiment of the invention, a television receiver includes an MPEG decoder. A data compression network quantizes a decoded and decompressed MPEG signal representing 8×8 image blocks prior to storage in frame memory, and reconstructs the image blocks when needed for an image display. A display device displays image information derived from the frame memory.
To maintain data integrity through the data reduction network, the network scans a pixel block and determines the range of image pixel values. The network performs a best fit between the actual range and a set of predetermined ranges. Using the selected predetermined range, the network accesses a quantization table containing quantization levels derived from the positive domain of the predetermined range. The quantization table is of sufficient scope to quantize DPCM values as described above. For this example, the bias value described above is the selected predetermined range. Since the selected predetermined range is generally less than the largest possible pixel value for the system (256 for eight bits), and because the quantization table (and corresponding dequantization table) contain only positive values, doubling quantization resolution. It is possible to increase quantization resolution by a factor of 32 if the smallest predetermined range is adequate for a given pixel block.
REFERENCES:
patent: 4422094 (1983-12-01), Lewis, Jr. et al.
patent: 4656500 (1987-04-01), Mori
patent: 5539865 (1996-07-01), Gentile
patent: 5764374 (1998-06-01), Seroussi et al.
“Video compression makes big gains”, by Peng H. Ang, Peter A. Ruetz, and David Auld—LSI Logic Corp., IEEE Spectrum—Oct., 1991, pp. 16-19.
“Fundamentals of Digital Image Processing”, by Anil K. Jain, pp. 476-498, Prentice-Hall International, Inc.
N.S.Jayant/P.Noll:“Digital Coding of Waveforms”, Prentice-Hall, pp. 188, 189, 191, 192, 283, and 286.
K.Pratt/H.G.Musmann:“Image Transmission Techniques-Advances in Electronics and Electron Physics,Suppl. 12”,1979,Academic Press, pp. 92-97.
Shuichi Matsumoto et al.:“120/140 MBIT/S Portable HDTV Codec and its Transmission Performance in a Field Trial Via Intelsat Satellite”,Signal Processing,Image Communication,vol. 4,No. 4/05, Aug. 1, 1992, pp. 359-377.
Shuichi Matsumoto et al.:“120/140 MBPS Intrafield DPCM System for Digital Transmission of HDTV Programs”, Signal Processing of HDTV, L'Aquila,Feb. 29-Mar. 2, 1988,No. Workshop 2,Feb. 29, 1988, Chiariglione L., pp. 509-516.
Saito M. et al.:“120 MBPS HDTV Codec and its Performance in a Field Trial”,Signal Processing of HDTV,2,TURIN,Aug. 30-Sep. 1, 1989,No. Workshop 3,Aug. 30, 1989, Chiariglione L., pp. 611-618.
L.S.Golding/P.M.Schultheiss:“Study of an Adaptive Quantizer”, Proceeding of the IEEE,vol. 55, No. 3,Mar. 1967,pp. 293-297.
PCT Search Report.
Beyers, Jr. Billy Wesley
Canfield Barth Alan
Lam Wai-man
Yu Haoping
Eriksen Guy H.
Kurdyla Ronald H.
Thomson Licensing S.A.
Tripoli Joseph S.
Wu Jingge
LandOfFree
Memory efficient compression apparatus in an image... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Memory efficient compression apparatus in an image..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Memory efficient compression apparatus in an image... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3012709