Quadtree-structured coding of color images and intra-coded images

• US 6,005,981 A
• Filed: 04/11/1996
• Issued: 12/21/1999
• Est. Priority Date: 04/11/1996
• Status: Expired due to Term

• Current Assignee: National Semiconductor Corporation (Texas Instruments, Inc.)
• Original Assignee: National Semiconductor Corporation (Texas Instruments, Inc.)
• Inventors: Ran, Xiaonong, Ng, Hak-Leong
• Primary Examiner(s): Au, Amelia
• Assistant Examiner(s): Johnson, Timothy M.
• Application Number:
  US08/631,652
• Time in Patent Office: 1,349 Days
• Field of Search: 382/166, 382/232, 382/240, 382/248, 382/281, 382/236, 382/244, 382/226, 358/430, 358/433, 358/432, 358/539, 348/398, 348/568, 348/408, 348/437, 348/438, 348/395, 348/393, 348/391, 348/396, 345/202, 341/79
• US Class Current: 382/240
• CPC Class Codes: G06T 9/40 Tree coding, e.g. quadtree, octree
• CPC Class Codes: H04N 1/41 Bandwidth or redundancy reduction by scanning H04N1/17 ; methods or arrangements for coding, decoding, compressing or decompressing digital video signals H04N19/00
• CPC Class Codes: H04N 19/107 between spatial and temporal predictive coding, e.g. picture refresh
• CPC Class Codes: H04N 19/61 in combination with predictive coding
• CPC Class Codes: H04N 19/96 Tree coding, e.g. quad-tree coding

Abstract

An encoding process is configurable to trade calculational complexity for bit rate while maintaining code format and image quality. A software encoder which implements the encoding process monitors bandwidth and processor utilization and when either is overtaxed, adjusts complexity to compensate. If further adjustment is required, encoded image quality is changed. Changing the search process or search window size used during searches for predictive blocks or changing a threshold which is compared to inter-code length before determining whether to calculate an intra-code changes complexity. An intra-coding method subtracts a base which depends on codes representing neighboring blocks from coefficients for a current block to determine a difference which is often small and easily compressed. Additionally, restricting a quadtree structure for one color component to a sub-tree of a quadtree structure for another other color component improves compression efficiency and reduces coding complexity.

Claims

1. A method for encoding an image formed by an array of pixels, wherein for each pixel, a combination including a first component and a second component indicates a color for the pixel, the method comprising:

determining a first quadtree structure for a first array which contains the first components of pixels in the image;

generating first code vectors, each first code vector indicating first components within a block in a partition of the first array associated with the first quadtree structure;

partitioning a second array which contains the second components into a second partition associated with the first quadtree structure;

merging blocks of the second partition to determine a third partition, the third partition being associated with a second quadtree structure that is a sub-tree of the first quadtree structure and is formed by removing branches from the first quadtree structure; and

generating second code vectors, each second code vector indicating the second components within a block in the third partition.

2. The method of claim 1, wherein the first components are luminances of the pixels, and the second components are chrominances of pixels.

3. The method of claim 1, wherein each first code vector comprises one or more coefficients from a Walsh transform of a corresponding block from the first partition, and each second code vectors comprises one or more coefficients from a Walsh transform of a corresponding block from the third partition.

4. The method of claim 1, further comprising:

generating a first code which indicates the second quadtree structure; and

generating a second code which indicates only the branches that must be added to the second quadtree structure to form the first quadtree structure.

5. The method of claim 1, wherein for each pixel, the combination indicating the color for the pixel further includes a third component, and wherein the method further comprises:

partitioning a third array which contains the third components into a fourth partition associated with the second quadtree structure;

merging blocks of the fourth partition to determine a fifth partition, the fifth partition being associated with a third quadtree structure that is a sub-tree of the second quadtree structure and is formed by removing branches from the second quadtree structure; and

generating third code vectors, each third code vector indicating the third components within a block in the fifth partition.

6. The method of claim 5, wherein the first components are luminances of the pixels, and the second and third components are chrominances of pixels.

7. The method of claim 5, further comprising:

generating a first code which indicates the third quadtree structure;

generating a second code which indicates the branches that when added to the third quadtree structure form the second quadtree structure; and

generating a third code which indicates the branches that when added to the second quadtree structure form the first quadtree structure.

8. A method for intra-coding an image, comprising:

partitioning the image into root blocks;

forming a quadtree partition of each root block, each quadtree partition including blocks of different sizes, wherein forming the quadtree partition of each root block, comprises;

determining a first quadtree partition for a first array representing the root block, wherein the first array contains first color components of pixels in the root block, the first quadtree partition being associated with a first quadtree structure;

partitioning a second array into a second partition associated with the first quadtree structure, wherein the second array contains second color components of pixels in the root block; and

merging blocks of the second partition to determine a third partition, the third partition being associated with a second quadtree structure that is a sub-tree of the first quadtree structure and is formed by removing branches from the first quadtree structure; and

encoding the blocks, wherein for each block in each quadtree partition, encoding comprises;

determining a vector indicating image content within the block;

selecting a base vector for the block, wherein if the block is encoded after at least one other block in the quadtree partition, the base vector depends upon vectors for one or more previously encoded blocks in the quadtree partition;

determining a difference between the vector and the base vector; and

encoding the difference.

9. A method for encoding an image, the method comprising:

applying a first merging procedure to a uniform partition of a first array to determine a first quadtree structure, wherein the first array contains first color components of pixels in the image;

partitioning a second array into a first partition associated with the first quadtree structure, wherein the second array contains second color components of the pixels in the image;

applying a second merging procedure to blocks that are in the first partition of the second array to determine a second partition of the second array, the second partition being associated with a second quadtree structure that is a sub-tree of the first quadtree structure and results from removing branches from the first quadtree structure; and

generating a code that indicates;

the second quadtree structure;

second color components for each block in the second partition of the second array;

the branches removed from the first quadtree structure to provide the second quadtree structure; and

first color components for each block in a partition associated with the first quadtree structure.

10. The method of claim 9, wherein the first and second merging procedures merge blocks in a partition only if merging the blocks increases error per pixel by less than a threshold amount.

11. The method of claim 9, further comprising:

partitioning a third array into a partition associated with the second quadtree structure, wherein the third array contains third color components of the pixels in the image; and

applying a third merging procedure to blocks that are in the partition of the third array to determine a partition associated with a third quadtree structure, the third quadtree structure being a sub-tree of the second quadtree structure and resulting from removing branches from the second quadtree structure, wherein

the code indicates the second quadtree structure by indicating the third quadtree structure and the branches removed from the second quadtree structure to form the third quadtree structure.

12. The method of claim 11, wherein:

the first merging procedure makes a determination to merge blocks based on the first color components of pixels in the image, without considering the second and third color components;

the second merging procedure makes a determination to merge blocks based on the second color components of pixels in the image, without considering the first and third color components; and

the third merging procedure makes a determination to merge blocks based on the third color components of pixels in the image, without considering the first and second color components.

13. The method of claim 9, wherein:

the first merging procedure makes a determination to merge blocks based on the first color components of pixels in the image, without considering the second color components; and

the second merging procedure makes a determination to merge blocks based on the second color components of pixels in the image, without considering the first color components.