Efficient Co-segmentation of Image Using Higher Order Clique

A new interactive image co segmentation algorithm using possibility estimation and higher order energy is proposed for extracting general foreground objects from a group of interrelated images. Our approach introduces the higher order cliques, energy into the co segmentation optimization process successfully. A region based likelihood estimation procedure is first performed to provide the primary knowledge for our higher order energy function. A new co segmentation energy function using higher order clique is developed, which can capably co segmentation energy function using higher order clique is developed, which can efficiently co segment the foreground objects with huge manifestation variations from a group of images in complex scenes. Both the quantitative and qualitative experimental results on representative datasets reveal that the accuracy of our co segmentation results is much higher than the stateof-the-art co segmentation methods.