Two papers accepted to CVPR 16: 1) a new method to localize the 6D Pose of an object; 2) a new way for multi-label segmentation, e.g. cells in an image, with the constraint that all connected foreground object should be convex.
Two papers accepted to CVPR 16: 1) a new method to localize the 6D Pose of an object; 2) a new way for multi-label segmentation, e.g. cells in an image, with the constraint that all connected foreground object should be convex.