Title :
Motion Compensated Frame Interpolation Based on Mixed Particle Filter and Genetic Algorithm Approach
Author :
Song, Haibin ; Men, Aidong ; Zheng, Xu
Author_Institution :
Sch. of Telecommun. Eng., Beijing Univ. of Posts & Telecommun., Beijing
Abstract :
We propose a new motion compensated frame interpolation scheme for video post-processing. Mixed particle filter (PF) and genetic algorithm (GA) approach is proposed to do the motion estimation, also a new fitness function considering object and subject visual quality evaluation is proposed. Beside, two new motion vector smoothing method are proposed to get a more consistency motion vector field. The simulation shows a high subjective and objective performance of the proposed scheme.
Keywords :
genetic algorithms; interpolation; motion estimation; particle filtering (numerical methods); smoothing methods; video signal processing; genetic algorithm; mixed particle filter; motion compensated frame interpolation; motion estimation; motion vector smoothing method; video post-processing; Decoding; Genetic algorithms; Genetic engineering; Interpolation; Motion estimation; Particle filters; Smoothing methods; Telecommunication computing; Video sequences; Videoconference; MCFI; genetic algorithm; motion estimation; particle filter;
Conference_Titel :
Natural Computation, 2008. ICNC '08. Fourth International Conference on
Conference_Location :
Jinan
Print_ISBN :
978-0-7695-3304-9
DOI :
10.1109/ICNC.2008.653
