Pseudo-fixed length image compression scheme

• US 6,285,796 B1
• Filed: 11/03/1997
• Issued: 09/04/2001
• Est. Priority Date: 11/03/1997
• Status: Expired due to Term

• Current Assignee: Intel Corporation
• Original Assignee: Intel Corporation
• Inventors: Yuan, Po, Keith, Michael, Acharya, Tinku, Goldstein, Judith A.
• Primary Examiner(s): Grant, II, Jerome
• Application Number:
  US08/963,335
• Time in Patent Office: 1,401 Days
• Field of Search: 358/261.2, 358/430, 348/397, 348/398, 382/239, 382/240, 382/238, 382/232, 382/251, 382/246
• US Class Current: 382/246
• CPC Class Codes: H04N 19/50 using predictive coding H04N19/61 takes precedence

Abstract

A method and apparatus for efficient image encoding in a predictive coding or other error image based image compression scheme, a pseudo-fixed length code is employed to convert certain calculated error values into codewords. The code is pseudo-fixed length since the length of codewords is fixed for a particular range of values and then increased for the next range. The length of codewords is not needed in decoding by choosing prefixes that uniquely distinguish codewords when packed together. As a result, efficient and quick image encoding/decoding is made possible.

Claims

1. A method comprising the steps of:

compressing image data using a pseudo-fixed length image compression technique, said pseudo-fixed length allocating more bits to words that represent information significant to human vision;

encoding into a pseudo-fixed length code a set of quantized error values, said pseudo-fixed length code encoding high frequency quantized error values and low frequency quantized error values with different fixed lengths, said low frequency values having a magnitude more visually significant than said high frequency values; and

transferring said compressed image data to a destination for retrieval and decompression.

2. A method according to claim 1 wherein the step of compressing includes the steps of:

generating a set of error values between said image data and predicted image data and quantizing said error values into a set of quantized error values such that said set of quantized error values has fewer values than said set of error values.

3. A method according to claim 2 wherein the step of transferring includes the steps of sequencing into words one or more pseudo-fixed length encoded quantized error values.

4. A method according to claim 1 further comprising the stop of decompressing said compressed image data into reconstructed image.

5. A method comprising:

compressing image data using a pseudo-fixed length image compression technique by;

generating a set of error values between said image data and predicted image data;

quantizing said error values into a set of quantized error values such that said set of quantized error values has fewer values than said set of error values;

encoding into a pseudo-fixed length code said set of quantized error values, said pseudo-fixed length code encoding high frequency quantized error values and low frequency quantized error values with different fixed lengths, said low frequency values having a magnitude more visually significant than said high frequency values; and

transferring said compressed image data to a destination for retrieval and decompression by;

determining the total number of quantized error values in said set of quantized error values;

determining the number of different desired frequency ranges for said quantized error values, said frequency ranges defining said low frequency quantized error values and said high frequency quantized error values; and

setting a fixed length for encoding quantized errors of a particular frequency range, a different fixed length being employed for each frequency range.

6. A method according to claim 5 wherein the number of different desired frequency range is 2.

7. A method according to claim 6 wherein the pseudo-fixed length used to encode low frequency quantized error values is half that of the pseudo-fixed length used to encode high frequency quantized error values.

8. An apparatus for image compression comprising:

an error image generator configured to generate image error values;

a pseudo fixed length encoder coupled to said error image of said encoder configured to encode said image error values into codewords of pseudo-fixed length, said pseudo-fixed length allocating more bits of words that represent information significant to human vision;

wherein said pseudo-fixed length encoder includes a look-up table addressable by said error image values, said look-up table providing therefrom a corresponding quantized error value.

9. An apparatus according to claim 8 further wherein said look-up table provides said pseudo-fixed length codewords for each quantized error value.

10. An apparatus according to claim 8 wherein said error image generator includes a predictive coding circuit.

11. An apparatus comprising:

an error image generator configured to generate image error values; and

a pseudo-fixed length encoder coupled to said error image of said encoder configured to encode said image error values into codewords of pseudo-fixed length an error image generator configured to generate image error values;

wherein said error image generator includes a predictive coding circuit including a difference circuit, a look-up table, and an error recovery circuit, said error recovery circuit providing one input to said difference circuit, said difference circuit having as a second input, pixel data, said difference circuit generating said image error values therefrom.

12. An apparatus according to claim 11 wherein said look-up table is indexed by the output of said difference and is configured to provide a recovered pixel value to said error recovery circuit.

13. An apparatus for image decompression comprising:

a pseudo-fixed length decoder configured to decode words into quantized error values from a look-up table addressable by said error image values, said look-up table providing therefrom a corresponding quantized error value.

14. An imaging system comprising:

a pseudo-fixed length image compression circuit for coupling to an imaging device configured to capture raw image data, said compression circuit configured to compress said raw image data, said pseudo-fixed length compression utilizing a set of look-up tables to perform said compression, the look-up tables are addressable said error image values, said look-up tables providing therefrom a corresponding quantized error value.

15. A system according to claim 14 wherein said stored compressed image data is output to data processing system.

16. A system according to claim 15 wherein said data processing system uncompresses said output image data utilizing a pseudo-fixed length decompression scheme.

17. A system according to claim 16 further comprising a display system, said display system configured to reorder said decompressed image data into a visually recognizable image.

18. A method comprising:

compressing image data using a pseudo-fixed length image compression technique;

transferring said compressed image data to a destination for retrieval and decompression;

generating a set of error values between said image data and predicted image data;

encoding into a pseudo-fixed length code said set of quantized error values, said pseudo-fixed length code encoding high frequency quantized error values and low frequency quantized error values with different fixed lengths, said low frequency values having a magnitude more visually significant than said high frequency values; and

quantizing said error values into a set of quantized error values such that said set of quantized error values has fewer values than said set of error values.

19. A method comprising:

compressing image data using a pseudo-fixed length image compression technique;

transferring said compressed image data to a destination for retrieval and decompression;

generating a set of error values between said image data and predicted image data;

quantizing said error values into a set of quantized error values such that said set of quantized error values has fewer values than said set of error values;

encoding into a pseudo-fixed length code said set of quantized error values, said pseudo-fixed length code encoding high frequency quantized error values and low frequency quantized error values with different fixed lengths, said low frequency values having a magnitude more visually significant than said high frequency values; and

determining the total number of quantized error values in said set of quantized error values.

20. An apparatus for image compression comprising:

an error image generator configured to generate image error values;

a pseudo fixed length encoder coupled to said error image of said encoder configured to encode said image error values into codewords of pseudo-fixed length, said pseudo-fixed length allocating more bits to words that represent information significant to human vision; and

a packing circuit configured to pack said pseudo-fixed length codewords for transport outside said apparatus.