Singular value decomposition coding and decoding apparatuses
US5615288A · kind A · utility
20Cited by
0References
8Claims
0Family size
Assignee
Inventors
Key dates
| Filing date | Apr 25, 1994 |
| Grant date | Mar 25, 1997 |
| Priority date | — |
| Expiry date | Apr 25, 2014 |
Classification
- Technology area (CPC G)Physics
- CPC primaryG06T9/001
- WIPO fieldComputer technology
- WIPO sectorElectrical engineering
Abstract
An input image is divided into blocks each represented by n.times.n matrix X, and then X.sup.T X is calculated. Singular values of X that are positive square roots of eigenvalues of X.sup.T X and first singular vectors that are normalized eigenvectors of X.sup.T X are calculated by, for instance, the Jacobi method. Second singular vectors that are normalized eigenvectors of XX.sup.T are calculated analytically using the singular values and the first singular vectors. The singular values and the first and second singular vectors thus calculated are coded.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.