Method for motion-compensated frame-to-frame prediction coding
US4816906A · kind A · utility
Assignee
Inventors
Key dates
| Filing date | Aug 31, 1987 |
| Grant date | Mar 28, 1989 |
| Priority date | — |
| Expiry date | Aug 31, 2007 |
Classification
- Technology area (CPC H)Electricity
- CPC primaryH04N19/61
- WIPO fieldAudio-visual technology
- WIPO sectorElectrical engineering
Abstract
A process of frame-to-frame prediction coding in which a frame is broken down into a regular block scanning field of n x n pixels and then prior to each prediction, up to four prediction methods, namely (1) block-by-block displacement vector calculation, (2) object-related displacement vector calculation, (3) nullification of the displacement, and (4) nullification of the prediction, are monitored as to the data rate to be transmitted, and the prediction method that leads to the smallest prediction errors is selected for the transmission.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.