Video saliency-detection using custom spatiotemporal fusion method

There have been several researches done in the field of image saliency but not as much video saliency. In order to increase precision and accuracy during compression, reduce coding complexity time consumption along with memory allocation problems our proposed solution. It is a modified high-definition compression (HEVC) pixel based consistent spatiotemporal diffusion temporal uniformity. involves taking apart into groups frames, computing colour saliency, integrate fusion, fusion conducted then information guides process for mapping help permutation matrix. The solution tested on publicly available extensive dataset five global valuation metrics compared other state-of-the-art detection methods. results display overall best performance amongst all candidates.