Method of compressing image signals using vector quantization
First Claim
1. A method of compressing image signals by:
- (i) dividing original image signals representing a two-dimensional image into blocks, each of said blocks being composed of a plurality of samples,(ii) converting the original image signals at respective picture elements in each of the blocks into an n-valued system where each of the plurality of samples has one of the n-values and, where n denotes an integer of 2 or larger,(iii) calculating mean values m1, m2, . . . , mn of the original image signals at the respective picture element in the block that have the common n-valued signals, and(iv) encoding the mean values m1, m2, . . . , mn and the n-valued signals,wherein the improvement comprises the steps of;
(a) carrying out vector quantization of said mean values m1, m2, . . . mn and/or said n-valued signals in each of said blocks.
2 Assignments
0 Petitions
Accused Products
Abstract
A method of compressing image signals comprises the steps of dividing original image signals representing a two-dimensional image into blocks each of which is composed of a plurality of samples, converting the original image signals at respective picture elements in each of the blocks into n-valued system where n denotes an integer of 2 or larger, calculating means values m1, m2, . . . , mn of the original image signals at the respective picture elements in the block that have the common n-valued signals, and encoding the mean values m1, m2, . . . , mn and the n-valued signals. The mean values m1, m2, . . . , mn and/or the n-valued signals in each of the blocks are then subjected to vector quantization.
46 Citations
3 Claims
-
1. A method of compressing image signals by:
-
(i) dividing original image signals representing a two-dimensional image into blocks, each of said blocks being composed of a plurality of samples, (ii) converting the original image signals at respective picture elements in each of the blocks into an n-valued system where each of the plurality of samples has one of the n-values and, where n denotes an integer of 2 or larger, (iii) calculating mean values m1, m2, . . . , mn of the original image signals at the respective picture element in the block that have the common n-valued signals, and (iv) encoding the mean values m1, m2, . . . , mn and the n-valued signals, wherein the improvement comprises the steps of; (a) carrying out vector quantization of said mean values m1, m2, . . . mn and/or said n-valued signals in each of said blocks. - View Dependent Claims (2, 3)
-
Specification