Method and apparatus for coding motion pictures

• US 5,121,191 A
• Filed: 03/15/1991
• Issued: 06/09/1992
• Est. Priority Date: 03/15/1991
• Status: Expired due to Term

• Current Assignee: Aware Incorporated (Crowley Holdings Preferred LLC)
• Original Assignee: Aware Incorporated (Crowley Holdings Preferred LLC)
• Inventors: Devine, Johnathan D., Huffman, John C., Cassereau, Philippe M., Zettler, William R. Jr.
• Primary Examiner(s): Britton, Howard W.
• Application Number:
  US07/669,773
• Time in Patent Office: 452 Days
• Field of Search: 358/11, 358/12, 358/13, 358/105, 358/133, 358/135, 358/136, 358/339
• US Class Current: 348/443
• CPC Class Codes: H04N 11/042 Codec means
• CPC Class Codes: H04N 19/10 using adaptive coding
• CPC Class Codes: H04N 19/115 Selection of the code volume for a coding unit prior to coding
• CPC Class Codes: H04N 19/146 Data rate or code amount at the encoder output
• CPC Class Codes: H04N 19/186 the unit being a colour or a chrominance component
• CPC Class Codes: H04N 19/62 by frequency transforming in three dimensions H04N19/63 takes precedence
• CPC Class Codes: H04N 19/63 using sub-band based transform, e.g. wavelets

Abstract

A method and apparatus for compressing motion pictures is disclosed. The frames of the motion picture are divided into adjacent groups. Each group is treated as a three-dimensional image. The three-dimensional image is then filtered via a three-dimensional FIR filter to generate three-dimensional component images that are more efficiently quantized. The degree of quantization of each component image is determined in part by the spatial frequencies represented by the component image in question. For motion pictures derrived from interlaced scanning devices, the quantization of specific component images is altered to prevent artifacts.

Claims

1. An apparatus for compressing a motion picture comprising a sequence of frames, each said frame comprising a two-dimensional image representing said motion picture at one instance in time, said apparatus comprising:

means for receiving a signal specifying a compression ratio for said motion picture;

means for receiving signals specifying said frames;

means for storing a predetermined number, N_t of consecutively received said frames;

three-dimensional filtering means for filtering said stored frames to generate a plurality of three-dimensional component images, said plurality of three-dimensional component images comprising a low-frequency component image and one or more high-frequency component images, each said component image comprising a plurality of pixel values; and

quantization means for approximating each pixel values in at least one of said component images by an approximation comprising one of a predetermined set of integer values, the precision of said integers being related to the statistical distribution of the pixel values in said component image and said received compression ratio; and

means for generating output signals specifying said pixel approximations.

2. The motion picture compression apparatus of claim 1 wherein said quantization means comprises means for approximating the pixels of said low-frequency image to a precision greater than the statistically optimum precision determined by said statistical distributions of said pixels in said component images and said received compression ratio;

and means for approximating said pixels of at least one of said high frequency component images to a precision less than the statistically optimum precision determined by the statistical distributions of said pixels in said component images and said received compression ratio.

3. The motion picture compression apparatus of claim 1 wherein said frames of said motion picture comprise interlaced frames, every other said frame being offset by one half the horizontal line spacing of said frames, wherein said receiving means comprises means for combining each pair of consecutively received frames to form a single frame having twice the number of horizontal lines, wherein said three-dimensional filtering means comprises means for generating at least one high frequency component image emphasizing horizontal edges in said frames of said motion picture, and wherein said quantization means comprises means for approximating the pixels of said high frequency component image emphasizing horizontal edges to a precision greater than said statistically optimum precision determined by the statistical distribution of said pixels in said high-frequency component image emphasizing horizontal edges.

4. The motion picture compression apparatus of claim 1 wherein each said frame of said motion picture comprises a luminance frame and first and second chrominance frames, wherein said three dimensional filtering means further comprises first and second three-dimensional chrominance filtering means, said first three-dimensional chrominance filtering means comprising means for filtering N_t of said first chrominance frames to generate a first set of three-dimensional filtered chrominance component images, said second three-dimensional chrominance filtering means comprising means for filtering the N_t said second chrominance frames corresponding to N_t of said first chrominance frames filtered by said first three-dimensional chrominance filtering means to generate a second set of three-dimensional filtered chrominance component images, and wherein said quantization means further comprises vector quantization means for approximating corresponding pixels in said first and second three-dimensional filtered chrominance component images by vectors of a predetermined precision.

5. The motion picture compression apparatus of claim 1 wherein said receiving means comprises:

two-dimensional filtering means for filtering each of said received frames to generate a two-dimensional low-frequency component image and one or more two-dimensional high frequency component images, each said component image comprising a plurality of pixel values, wherein

said storing means comprises a plurality of frame buffers, each said frame buffer comprising means for storing said component images generated from one of said consecutively received frames, and wherein

said three-dimensional filtering means comprises a one-dimensional filtering means for combining corresponding pixels in said frame buffers.

6. The motion picture compression apparatus of claim 5 wherein said two-dimensional filtering means comprises a one-dimensional conjugate mirror filter bank.

7. An apparatus for decompressing a compressed motion picture, said compressed motion picture comprising a sequence of approximations for the pixels of one or more three-dimensional images, said decompression apparatus comprising:

means for receiving a signal specifying said pixel approximations;

means for generating one or more three-dimensional component images from said received pixel approximations including means for generating blank pixel values for any said pixel value not received;

inverse three-dimensional filtering means for combining said three-dimensional component images to generate a plurality of two-dimensional luminance picture frames;

means for generating output signals specifying said generated luminance picture frames;

8. The motion picture decompression apparatus of claim 7 wherein said pixel approximations comprise a plurality of two dimensional vector values, the components of each said vector value specifying first and second chrominance pixel values, wherein said means for generating three-dimensional component images comprises means for assembling said first and second chrominance pixel values into first and second sets of chrominance component images, and wherein said inverse three-dimensional filtering means further comprises means for combining said first and second sets of chrominance component images into first and second two-dimensional chrominance picture frames, and wherein said output signal generating means further comprises means for outputting said chrominance picture frames.

9. A method for compressing a motion picture to a given compression ratio, said motion picture comprising a sequence of frames, each said frame comprising a two-dimensional image representing said motion picture at one instance in time, said apparatus comprising:

receiving signals specifying said frames;

storing a predetermined number N_t of consecutively received said frames;

filtering said stored frames to generate a plurality of three-dimensional component images, said plurality of three-dimensional component images comprising a low-frequency component image and one or more high-frequency component images, each said component image comprising a plurality of pixel values; and

approximating each pixel values in at least one of said component images by an approximation comprising one of a predetermined set of integer values, the precision of said integers being related to the statistical distribution of the pixel values in said component image and said compression ratio; and

generating output signals specifying said pixel approximations.

10. The method of claim 9 wherein said approximating step comprises approximating the pixels of said low-frequency image to a precision greater than the statistically optimum precision determined by said statistical distributions of said pixels in said component images and said compression ratio;

and approximating said pixels of at least one of said high frequency component images to a precision less than the statistically optimum precision determined by the statistical distributions of said pixels in said component images and said compression ratio.

11. The method claim 9 wherein said frames of said motion picture comprise interlaced frames, every other said frame being offset by one half the horizontal line spacing of said frames, wherein said receiving step comprises combining each pair of consecutively received frames to form a single frame having twice the number of horizontal lines, wherein said three-dimensional filtering step comprises generating at least one high frequency component image emphasizing horizontal edges, and wherein said approximating step comprises approximating the pixels of said high frequency component image emphasizing horizontal edges to a precision greater than said statistically optimum precision determined by the statistical distribution of said pixels in said high-frequency component image emphasizing horizontal edges.

12. The method of claim 9 wherein each said frame of said motion picture comprises a luminance frame and first and second chrominance frames, wherein said filtering step further comprises filtering N_t of said first chrominance frames to generate a first set of three-dimensional filtered chrominance component images, filtering the N_t said second chrominance frames corresponding to said N_t said first chrominance frames to generate a second set of three-dimensional filtered chrominance component images, and wherein said approximating step approximating corresponding pixels in said first and second three-dimensional filtered chrominance component images by vectors of a predetermined precision.

13. The method claim 9 wherein said receiving step comprises:

filtering each of said received frames to generate a two-dimensional low-frequency component image and one or more two-dimensional high frequency component images, each said component image comprising a plurality of pixel values, wherein

said storing step comprises storing said two-dimensional component images in a plurality of frame buffers, each said frame buffer comprising means for storing said component images generated from one of said consecutively received frames, and wherein

said filtering step comprises combining corresponding pixels in said frame buffers.

14. The method of claim 13 wherein said two-dimensional filtering step comprises filtering said pixels with a one-dimensional conjugate mirror filter bank.

15. A recording of a motion picture comprising a plurality of ordered frames, said recording comprising means for storing a plurality of digital values, said digital values being divided into groups, each said group comprising approximations for the pixels of one or more component images, said component images specifying a three-dimensional image, each said three dimensional image comprising information specifying two or more frames of said motion picture, said component images comprising a low-frequency component image and one or more high-frequency component images.

16. The recording of claim 15 wherein each said group further comprises digital values specifying two-dimensional chrominance vectors, said chrominance vectors specifying first and second three-dimensional chrominance component images, said first and second chrominance component sub-images specifying a first and second three-dimensional chrominance images.

17. The recording of claim 15 wherein each said group comprises a high frequency component image emphasizing horizontal image features in the frames of said motion picture.