The Technology Of Video Scaling Algorithm Based On Temporal And Spatial Saliency

On the condition of different monitoring systems and display equipments, the image andvideo should be scaled to meet the need. Conventional image scaling technology cannot achievegood results, due to missing content, image distortion these phenomena. The technology ofdetection makes it possible to solve these problems. In this paper, we proposed a video scalingalgorithm based on spatio-temporal saliency detection.Firstly, the method improved the Itti model by replacing direction with texture information,including extracting the image color, brightness, texture information and fusing the three maps intospatial saliency map, which can reflect the spatial significant regional information; Secondly, forthe video sequence, the adaptive kernel density estimation is used to calculate the time, saliencymap in order to extract the moving object information effectively; then, these two saliency mapsare weighted combined into spatio-temporal saliency map by dynamic integration; Finally thesaliency and non saliency regions are scaled respectively by fitted function and zoom parametersin order to complete the image and video scaling.The algorithm is tested in videos which have obvious movements. Compared with Itti andGBVS, the method can accurately detect the salient areas in images and videos. The scaling ofimage and video ensures the significant zone information and contents completion. Theexperimental results show that the proposed method has higher efficiency and less time.