Method of calculating motion vectors
US6160917A · kind A · utility
Assignee
Inventor
Key dates
| Filing date | Aug 13, 1998 |
| Grant date | Dec 12, 2000 |
| Priority date | — |
| Expiry date | Aug 13, 2018 |
Classification
- Technology area (CPC H)Electricity
- CPC primaryH04N19/51
- WIPO fieldAudio-visual technology
- WIPO sectorElectrical engineering
Abstract
A method in which motion vectors relating to frames and fields are calculated from motion vectors for a digital picture signal wherein for motion vectors relating to frames, the method wherein PA1 motion vectors provided for fields are converted to motion vectors for frames, PA1 the motion vectors for frames are converted in such a way that the motion vectors relate exclusively to the previous frame, PA1 motion vectors possibly missing from a frame for one or more blocks are determined by filtering from the motion vectors of blocks of the same frame and/or blocks of the previous frame, a block row and a block column, in which the x components of the motion vectors have the smallest spread, and a block row and block column, in which the y components of the motion vectors have the smallest spread, are selected from the motion vectors thus preprocessed, PA1 the global motion vectors are calculated for each frame, in which the x components of the preprocessed motion vectors required for their calculation are taken exclusively from the block row and block column selected for these components, and in which the required y components of the preprocessed motion vectors are taken exclusively…
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.