A Method of Image Compression Based on Visual Salient Features
Hui Wang, Gang Liu, Hongchang Ke, Liufu Song · International Conference on Mechanic Automation and Control Engineering · 2012
For image compression at region of interest, especially those with motion features image compression at region of interest, a calculation model based on the visual saliency fusing into the motion features of is proposed. First the original features of the image are extracted, and then a salient image is generated to calculate the region of interest. Then a image compression method based on this model is proposed, region of interest is to do low compression ratio, or even uncompressed and the background region is to do high compression ratio. This method can rapidly calculate the region of interest, and has highly compression efficiency and visual effect.