As a key frame extraction method and system based on visual attention model

The present invention discloses a method and system for key frame extraction based on visual attention model. Extracting method comprising: in the airspace, the method of detecting a significant global contrast filter binomial coefficients, and the target region extracted by the adaptive threshold. Algorithm not only can better maintain the target area boundary significantly, but also significantly more uniform degree in the area. Then, in the time domain, which defines a significant degree of movement, to estimate the target motion by a homography matrix using key instead of the target for a significant degree detecting, after the fused data spatial saliency proposed energy function boundaries based the method for obtaining a significant extension target region domain as the bounding box. Finally, the method reduces the richness of the video by a significant target area, using a combination of an adaptive clustering method were shot online key frame extraction.