Abstraction for Static and Dynamic Images Based on Visual Saliency

Reliable estimation of visual perception reflects understanding of the meaningful structure in an image. The paper describes a complete abstraction framework for static and dynamic images that explicitly responds to this goal, stylizes the corresponding salient and non-salient part differently by an edge preserving filter at a time while keep harmonious transition between the two parts. First, this paper introduces an automatic salient object segmentation algorithm to distinguish salient regions, and it is a saliency computation based local spatial neighbors. Taking into account the actual needs, the paper provides an interactive technology, this can be convenient for specifying image salient structure information on purpose. Based on the generated salient information mask, the paper then uses single-scale an isotropic filter to process the salient part, and use multi-scale an isotropic filter to process the non-salient part so that can implement a strong abstraction effect. Proposed method generates a kind of image and video abstraction that it represents a preferable visual effect. Experiments show that our algorithm could get the desired effect for processing a certain number of images and videos.