Title :
A filtering method of motion vetcor fields available for digital vision chips
Author :
Ueyama, O.E. ; Hosoe, Shigeyuki ; Hosoe, Shigeyuki
Author_Institution :
Tsu City Coll., Japan
Abstract :
We report a noise reduction algorithm concerning about visual motion processing, which can be implemented with digital vision chips. The algorithm is based on the maximum a posteriori probability estimation for Markov random vector fields, where Cauchy distribution is used to reduce noise and to detect motion discontinuities (edges). Simulation results show that good estimate and their distinct discontinuities are obtained within a few steps of iterations.
Keywords :
Markov processes; digital filters; digital signal processing chips; maximum likelihood estimation; motion estimation; Cauchy distribution; Markov random vector field; digital vision chips; filtering method; maximum a posteriori probability estimation; motion discontinuity detection; motion edge; motion vector estimation; noise reduction algorithm; visual motion processing;
Conference_Titel :
SICE 2004 Annual Conference
Conference_Location :
Sapporo
Print_ISBN :
4-907764-22-7
