This paper advances prior work by proposing a joint learning framework to simultaneously identify the spatial and temporal extents of the action of interest in training videos. To get pixel-level localization results, our method uses dense traj

提出一种利用视觉显著性对图像进行分割的方法。首先提取图像的底层视觉特征，从局部显著性、全局显著性和稀少性3个方面计算各特征图像中各像素的视觉显著性，得到各特征显著图；对各特征显著图进行综合，生 成最终的综合显著图。然后对综合显著图进行阈值分割，得到二值图像，将二值图像与原始图像叠加，将前景和背景分离，得到图像分割结果-It presents a significant advantage of visual image segmentation method. First, extract

高斯处理视频并跟踪运动做前景背景分割，可用来做前背景分离，分离出运动物体，达到抠图的效果-Gauss video processing and tracking moving foreground background segmentation, can be used to do before the separation, separation of a moving object, to achieve the effect of Gauss matting video processin