Institute of Information Science, Academia Sinica

Video object cosegmentation concerns the task of segmenting the common object in a pair of video clips. We present a new algorithm that works on super-voxels to solve this task. The algorithm computes i) the intra-video relative motion derived from dense optical flow and ii) the inter-video co-features based on Gaussian mixture models. We show that, by integrating the intra-video and inter-video information, our algorithm is able to produce good video-object segmentation results.