Title :
Occam algorithms for computing visual motion
Author :
Schweitzer, Haim
Author_Institution :
Texas Univ., Dallas, Richardson, TX, USA
Abstract :
By drawing an analogy with machine learning, the author proposes to define visual motion as a predictor that can accurately predict future frames. Under this new definition, visual motion can be specified by a collection of image patches, each moving in a simple motion. An implementation with rectangular patches determined recursively by a binary decision tree is described. Experimental results on real video sequences verify the algorithm assumptions and show that motion in typical sequences can be accurately described in terms of a few parameters
Keywords :
Occam; computer vision; decision theory; learning (artificial intelligence); Occam algorithms; binary decision tree; image patches; machine learning; real video sequences; visual motion computing; Acceleration; Constraint optimization; Decision trees; Encoding; Image motion analysis; Machine learning; Machine learning algorithms; Motion estimation; Pixel; Video sequences;
Conference_Titel :
Computer Vision, 1993. Proceedings., Fourth International Conference on
Conference_Location :
Berlin
Print_ISBN :
0-8186-3870-2
DOI :
10.1109/ICCV.1993.378163
