OPTICAL-FLOW ESTIMATION AND MOVING OBJECT SEGMENTATION BASED ON MEDIAN RADIAL BASIS FUNCTION NETWORK

Various approaches have been proposed for simultaneous optical flow es timation and segmentation in image sequences, In this study, the movin g scene is decomposed into different regions with respect to their mot ion, by means of a pattern recognition scheme, The inputs of the propo sed scheme are the feature vectors representing still image and motion information. Each class corresponds to a moving object. The classifie r employed is the median radial basis function (MRBF) neural network. An error criterion function derived from the probability estimation th eory and expressed as a function of the moving scene model is used as the cost function. Each basis function is activated by a certain image region. Marginal median and median of the absolute deviations from th e median (MAD) estimators are employed for estimating the basis functi on parameters. The image regions associated with the basis functions a re merged by the output units in order to identify moving objects.