Video retrieval based on dynamics of color flows

Content based video retrieval is particularly challenging because the huge amount of data associated with videos complicates the extraction of salient information content descriptors. Commercials are a video category where a large part of the content depends on low level perceptual features such as colors and color dynamics. These are related to the evolution-in terms of shrinking, growth and translation"of colored regions along consecutive frames. Each colored region, during its evolution, defines a 3D volume: a color flow. In the paper, a system is presented that supports description of color flows based on 3D wavelet decomposition and retrieval of commercials based on color flow similarity.