Method to control the generated bit rate in MPEG-4 shape coding

• US 6,081,554 A
• Filed: 10/02/1998
• Issued: 06/27/2000
• Est. Priority Date: 10/02/1998
• Status: Expired due to Fees

• Current Assignee: Trustees Of Columbia University In The City Of New York (Columbia University)
• Original Assignee: Trustees Of Columbia University In The City Of New York (Columbia University)
• Inventors: Lee, Jae-Boom, Eleftheriadis, Alexandros
• Primary Examiner(s): Britton, Howard
• Assistant Examiner(s): DIEP, NHON THANH
• Application Number:
  US09/166,212
• Time in Patent Office: 634 Days
• Field of Search: 348/390, 348/404, 348/405, 348/419, 348/420, 348/421, 382/239, 382/243
• US Class Current: 375/240
• CPC Class Codes: G06T 9/20 Contour coding, e.g. using detection of edges
• CPC Class Codes: H04N 19/21 with binary alpha-plane coding for video objects, e.g. context-based arithmetic encoding [CAE]

Abstract

In a digital video compression system that includes both a real buffer of size B_max, and a smaller virtual buffer of size B_virtual, a method for controlling the generated bit rate of compressed video information to keep within the maximum buffer capacity is disclosed. The method includes the steps of (a) receiving blocks of digital information, (b) determining whether a current block can be compressed by one of one or more shortcuts and using the shortcut if possible; (c) if the current block cannot be compressed with any shortcut, determining whether the virtual buffer capacity will be exceeded if the current received block is compressed only by arithmetic coding and using only arithmetic coding if possible; (d) if the current block cannot be compressed with any shortcut and the virtual buffer capacity would be exceeded with only arithmetic coding, determining whether the virtual buffer capacity will be exceeded if the current block is compressed with both downsampling and arithmetic coding and using both downsampling and arithmetic coding if possible; and (e) if all else fails, compressing the current block with a default mode of compression even if the virtual buffer capacity is exceeded with such compression.

Claims

1. In a digital video compression system comprising a real buffer having a maximum buffer capacity and virtual buffer having a virtual buffer capacity, said virtual buffer capacity being less than said maximum buffer capacity, a method for controlling the generated bit rate of compressed video information to keep within said maximum buffer capacity comprising the steps of:

(a) receiving one or more blocks of digital information, each block representing at least one identifiable characteristicfor a portion of a frame of video;

(b) determining whether a current received block can be compressed by one of one or more shortcuts, and if so, compressing said current received block with one of said shortcuts;

(c) if said current received block cannot be compressed with said shortcuts, determining whether said virtual buffer capacity will be exceeded if said current received block is compressed only by arithmetic coding, and if not, compressing said current received block with only arithmetic coding;

(d) if said current received block cannot be compressed with said shortcuts and if said virtual buffer capacity would be exceeded with only arithmetic coding, determining whether said virtual buffer capacity will be exceeded if said current received block is compressed with both downsampling and arithmetic coding, and if not, compressing said current received block with both said downsampling and said arithmetic coding; and

(e) if said current received block cannot be compressed with said shortcuts and if said virtual buffer capacity would be exceeded with only arithmetic coding or with both downsampling and arithmetic coding, compressing said current received block with a default mode of compression even if said virtual buffer capacity is exceeded with such compression.

2. The method of claim 1, wherein said digital video compression system is an MPEG-4 compliant shape code, said one or more blocks of digital information arc binary alpha blocks, and said current received block is a current binary alpha block.

3. The method of claim 2, wherein said at least one identifiable characteristic is binary shape.

4. The method of claim 2, wherein said one or more shortcuts comprise a lossless compression mode representing an entirely white binary alpha block, a lossless compression mode representing an entirely black binary alpha block, and a lossless compression mode representing no change in an binary alpha block.

5. The method of claim 2, wherein said arithmetic coding comprises content-based arithmetic encoding.

6. The method of claim 5, wherein said step of content-based arithmetic encoding comprises the step of selecting between intra-coding of said binary alpha block, inter-coding of said binary alpha block, intra-coding of a transpose of said binary alpha block, and inter-coding of a transpose of said binary alpha block.

7. The method of claim 6, wherein said selecting step comprises choosing the content-based arithmetic encoding yielding the fewest bits.

8. The method of claim 2, wherein downsampling is size compression at a ratio of 1:

2, and said step of compressing said current received block with both said downsampling and arithmetic coding comprises size compressing said current binary alpha block to 1/2 it original size and then selecting between intra-coding of said size compressed binary alpha block, inter-coding of said size compressed binary alpha block, intra-coding of a transpose of said size compressed binary alpha block, and inter-coding of a transpose of said size compressed binary alpha block.

9. The method of claim 8, further comprising the step of determining, if said current binary alpha block cannot be compressed with said shortcuts and if said virtual buffer capacity would be exceeded with either only arithmetic coding or with both size compression at a ratio of 1:

2 and with arithmetic coding, size compressing said current binary alpha block to 1/4 it original size and then selecting between intra-coding of said size compressed binary alpha block, inter-coding of said size compressed binary alpha block, intra-coding of a transpose of said size compressed binary alpha block, and inter-coding of a transpose of said size compressed binary alpha block.

10. The method of claim 9, wherein said default mode of compression comprises size compressing said current binary alpha block to 1/4 it original size and then selecting, from intra-coding of said size compressed binary alpha block, inter-coding of said size compressed binary alpha block, intra-coding of a transpose of said size compressed binary alpha block, and inter-coding of a transpose of said size compressed binary alpha block, the lossless compression mode yielding the fewest bits.

11. A method for compressing digital video information by downsampling and arithmetic encoding while controlling the generated bit rate of compressed video information to keep within a maximum buffer capacity by employing a virtual buffer having a capacity being less than said maximum buffer capacity, comprising the steps of:

(a) receiving one or more blocks of digital information, each block representing at least one identifiable characteristic for a portion of a frame of video;

(b) determining whether a current received block can be compressed by one of one or more shortcuts, and if so, compressing said current received block with one of said shortcuts;

(c) if said current received block cannot be compressed with said shortcuts;

determining whether said virtual buffer capacity will be exceeded if said current received block is compressed only by arithmetic coding, and if not, compressing said current received block with only arithmetic coding;

(d) if said current received block cannot be compressed with said shortcuts and if said virtual buffer capacity would be exceeded with only arithmetic coding, determining whether said virtual buffer capacity will be exceeded if said current received block is compressed with both downsampling and arithmetic coding, and if not, compressing said current received block with both said downsampling and said arithmetic coding; and

(e) if said current received block cannot be compressed with said shortcuts and if said virtual buffer capacity would be exceeded with only arithmetic coding or with both downsampling and arithmetic coding, compressing said current received block with a default mode of compression even if said virtual buffer capacity is exceeded with such compression.

12. The method of claim 11, wherein said digital video compression system is an MPEG-4 compliant shape code, said one or more blocks of digital information are binary alpha blocks, said current received block is a current binary alpha block, and said at least one identifiable characteristic is binary shape.

13. The method of claim 12, wherein said one or more shortcuts comprise a lossless compression mode representing an entirely white binary alpha block, a lossless compression mode representing an entirely black binary alpha block, and a lossless compression mode representing no change in an binary alpha block.

14. The method of claim 12, wherein said arithmetic coding comprises content-based arithmetic encoding.

15. The method of claim 14, wherein said step of content-based arithmetic encoding comprises the step of selecting between intra-coding of said binary alpha block, inter-coding of said binary alpha block, intra-coding of a transpose of said binary alpha block, and inter-coding of a transpose of said binary alpha block.

16. The method of claim 15, wherein said selecting step comprises choosing the content-based arithmetic encoding yielding the fewest bits.

17. The method of claim 12, wherein downsampling is size compression at a ratio of 1:

2, and said step of compressing said current received block with both said downsampling and arithmetic coding comprises size compressing said current binary alpha block to 1/2 it original size and then selecting between intra-coding of said size compressed binary alpha block, inter-coding of said size compressed binary alpha block, intra-coding of a transpose of said size compressed binary alpha block, and inter-coding of a transpose of said size compressed binary alpha block.

18. eThe method of claim 17, further comprising the step of determining, if said current binary alpha block cannot be compressed with said shortcuts and if said virtual buffer capacity would be exceeded with either only arithmetic coding or with both size compression at a ratio of 1:

2 and with arithmetic coding, size compressing said current binary alpha block to 1/4 it original size and then selecting between intra-coding of said size compressed binary alpha block, inter-coding of said size compressed binary alpha block, intra-coding of a transpose of said size compressed binary alpha block, and inter-coding of a transpose of said size compressed binary alpha block.

19. The method of claim 18, wherein said default mode of compression comprises size compressing said current binary alpha block to 1/4 it original size and then selecting, from intra-coding of said size compressed binary alpha block, inter-coding of said size compressed binary alpha block, intra-coding of a transpose of said size compressed binary alpha block, and inter-coding of a transpose of said size compressed binary alpha block, the lossless compression mode yielding the fewest bits.

20. A method for compressing digital video information by downsampling and arithmetic encoding while controlling the generated bit rate of compressed video information to keep within a maximum buffer capacity, comprising the steps of:

(a) receiving one or more blocks of digital information, each block representing at least one identifiable characteristic for a portion of a frame of video;

(b) determining whether a current received block can be compressed by one of one or more shortcuts, and if so, compressing said current received block with one of said shortcuts;

(c) if said current received block cannot be compressed with said shortcuts, determining a measure of distortion if said current received block is only arithmetic encoded or first downsampled and then arithmetic encoded and

(1) compressing said current received block by first downsampling and then arithmetic encoding if such encoding yields acceptable distortion and if said maximum buffer capacity is not exceeded;

(2) if downsampling and then arithmetic encoding does not yield acceptable distortion, compressing said current received block with only arithmetic encoding if said maximum buffer capacity is not exceeded; and

(3) if said maximum buffer capacity is exceeded, recompressing a previous block in order to reduce buffer occupancy.