RECOGNITION OF DYNAMIC VIDEO CONTENTS BASED ON MOTION TEXTURE STATISTICAL MODELS
Tomás Crivelli, Bruno Cernuschi-Frías, Patrick Bouthémy, Jianfeng Yao · 2008
IRMAR/Univ. of Rennes 1, Campus de Beaulieu, 35042 Rennes Cedex, [email protected]: Motion analysis, Markov random fields, image content classi fication, dynamic textures.Abstract: The aim of this work is to model, learn and recognize, dynamic contents in video sequences, displayed mostlyby natural scene elements, such as rivers, smoke, moving foliage, fire, etc. We adopt the mixed-stateMarkovrandom fields modeling recently introduced to represent the so-called motion textures. The approach consistsin describing the spatial distribution of some motion measurements which exhibit values of two types: adiscrete component related to the absence of motion and a continuous part for measurements different fromzero. Based on this, we present a method for recognition and classification of real motion textures using thegenerative statistical models that can be learned for each motion texture class. Experiments on sequences fromthe DynTex dynamic texture database demonstrate the performance of this novel approach.