An effective object segmentation method for 2D to 3D conversion

In this thesis, I propose the semiautomatic object segmentation technique for 2D to 3D conversion. I used the GrabCut algorithm for the semiautomatic object segmentation technique method, and the GrabCut algorithm has a disadvantage that requires the user to manually input data about primary Trimap. There are also a good number of computations, although its segmentation result is relatively more accurate. The suggested method is to perform object segmentation automatically in the next frame after using the result from GrabCut for the first frame in image sequence, and then mat the object region for easy 2D to 3D conversion by using the snake algorithm. You can perform object segmentation more efficiently in image sequence by using this method.