Back to results
US 6,608,865 B1Granted
Coding method for video signal based on the correlation between the edge direction and the distribution of the DCT coefficients
AssigneeTexas Instruments Incorporated
InventorYuji Itoh
Life of this patent
Estimating the term…
Abstract
The objective of the invention is to make high-efficiency compressed coding for images. In coding of video signals in this application example, the edge is extracted in a frame of the input image in block 10. In block 11, classification is made into blocks where edge is present and blocks free of edge. In block 12, for the block where edge is present, based on the correlation between the edge direction and the distribution of the DCT coefficients, the optimum scanning order is determined and coding is performed by the scanning order.
Claims9
A method of coding individual frames in a video signal comprising the steps of:
- performing an edge extraction upon the entire frame;
- coding any extracted edges by starting coordinate, one of a predetermined plurality of directions and length;
- dividing the frame into a plurality of blocks;
- classifying each of said plurality of blocks into one of a plurality of categories, the plurality of categories including a first category of blocks including no extracted edges, a plurality of second categories of blocks having one extracted edge and a corresponding one of said predetermined plurality of directions and a third category of blocks including more than one extracted edge;
- converting each block from spatial domain pixels into the DCT coefficients via a Discrete Cosine Transform (DCT);
- providing a plurality of DCT coefficient scanning orders including a default zig-zag scanning order and a plurality of directional scanning orders, each directional scanning order corresponding to one of said predetermined plurality of directions and having an order corresponding to an expected distribution of DCT coefficients for said corresponding direction;
- and
- variable length coding said DCT coefficients of each block employing a corresponding one of said DCT coefficient scanning orders, whereby blocks classified in said first and said third categories are scanned with said default zig-zag scanning order and blocks classified in one of said second categories are scanned with a directional scanning order for said corresponding direction.
5 more elements
A method of coding individual frames in a video signal comprising the steps of:
- performing an edge extraction upon the entire frame;
- following said step of performing edge extraction coding any extracted edges by starting coordinate, one of a predetermined plurality of directions and length;
- dividing the frame into a plurality of blocks;
- classifying each of said plurality of blocks into blocks including no extracted edges and blocks having at least one extracted edge;
- converting each block including no extracted edges from spatial domain pixels into the DCT coefficients via a Discrete Cosine Transform (DCT);
- dividing each block having at least one extracted edge into a plurality of subblocks;
- converting each subblock from spatial domain pixels into the DCT coefficients via a Discrete Cosine Transform (DCT);
- providing a plurality of DCT coefficient scanning orders including a default zig-zag scanning order and a plurality of directional scanning orders, each directional scanning order corresponding to one of said predetermined plurality of directions and having an order corresponding to an expected distribution of DCT coefficients for said corresponding direction;
- classifying each of said plurality of subblocks into one of a plurality of categories, the plurality of categories including a first category of subblocks including no extracted edges, a plurality of second categories of subblocks having one extracted edge and a corresponding one of said predetermined plurality of directions and a third category of subblocks including more than one extracted edge;
- variable length coding said DCT coefficients of each block and each subblock, said step of variable length coding said DCT coefficients of each block and each subblock
- employs said default zig-zag scanning order for blocks having no extracted edges,
- employs said default zig-zag scanning order for subblocks classified in said first and said third categories, and
- employs a directional scanning order for subblocks classified in one of said second categories according to said corresponding direction.
10 more elements
Dependencies are inferred from the claim text.