Digital nonalgorithmic method of compressing data which represents addresses and values
US5271071A · kind A · utility
Assignee
Inventor
Key dates
| Filing date | Jun 21, 1991 |
| Grant date | Dec 14, 1993 |
| Priority date | — |
| Expiry date | Jun 21, 2011 |
Classification
- Technology area (CPC H)Electricity
- CPC primaryH04N19/98
- WIPO fieldAudio-visual technology
- WIPO sectorElectrical engineering
Abstract
There is provided a digital, non-algorithmic method of compressing data. Data representing the addresses and values of pixels in an array are compressed. A first matrix is produced. The addresses of the first matrix are isomorphic with the range of possible pixel values, with one address corresponding to one pixel value. Same-valued pixels in the array are counted and each count is entered or tallied in the address of the first matrix corresponding to the value. One second matrix is generated for each pixel value actually present in the array. The addresses of the array and the second matrices are isomorphic. A binary "1" is placed in each address of each second matrix which corresponds to the address of the array having a pixel with the value to which the second matrix corresponds. The generated matrices represent very high data compression of 99% or higher. The first and second matrices may be transmitted, following which the array may be reconstructed therefrom. Because of the digital, non-algorithmic character of the foregoing, compression, transmission and decompression are essentially lossless and are conveniently subject to error correction. Residual mathematics is used to c…
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.