2D Visual Object Segmentation

In computer vision, 2D visual object segmentation is a key problem. In this thesis, we focused on the 2D visual object segmentation for the two 2D input data of computer vison: image and video. So our research included two parts: Image Object Segmentation and Video Object Segmentation,, Image Object SegmentationWe used active statistical object model, including Active Shape Models(ASM) and Active Appearance Models(AAM), for face image object segmentation. We proposed weighted ASM, shape subspace optimized ASM and subspace optimized AAM.- Active Shape Models use the orthogonal projection to the shape subspace, neglecting the information during the search procedure. So, a Weighted Active Shape Models(WASM) is proposed, using the search information. The search information is based on the local appearance model of ASM to determine how well the searching shape match models derived from the training set. It is used to project the intermediate search shape to the shape subspace. Compared with ASM's orthogonal projection, the weighted projection can drag the search out of local minima to be more accurate and more robust. Experiments have been done to show the ability of this method to align shapes. Active Shape Models(ASM) is composed of two parts: the ASM shape subspace model and the ASM search. While these two parts are closely interrelated and the performance of ASM depends on both of them, existing efforts treated them separately and had not considered how to optimize them overall. In their methods, the ASM shape subspace model was trained without considering the search procedure and the ASM search was performed using this shape subspace model without considering how it was trained. So we proposed an approach to optimize the shape model while considering the ASM search. We first perform an error analysis of ASM, and then to minimize the ASM error we propose an approach which optimizes the shape model according to the ASM search. For the ASM error analysis, we decomposed the ASM error into two parts, which are introduced by the ASM model and the ASM search respectively. With this decomposition, we prove that the optimal results of ASM can be achieved only by optimizing both of them jointly rather than separately. Furthermore, based on this error decomposition, we develop a method to find the optima! ASM shape model according to the ASM search by considering both the two decomposed errors. Experimental results demonstrate that our method can find the optimal ASM shape model rapidly and improve the performance of ASM significantly.- Similar to ASM, Active Appearance Models(AAM) is also composed of two parts: the AAM subspace model and the AAM search. While these two parts are closely correlated, existing efforts treated them separately and had not considered how to optimize them overall. In this paper, an approach is proposed to optimize the subspace model while considering the search procedure. We first perform a subspace error analysis, and then to minimize the AAM error we propose an approach which optimizes the subspace model according to the search procedure. For the subspace error analysis, we decomposed the subspace error into two parts, which are introduced by the subspace model and the search procedure respectively. This decomposition shows that the optimal results of AAM can be achieved only by optimizing both of them jointly rather than separately. Furthermore,based on this error decomposition, we develop a method to find the optimal subspace model according to the search procedure by considering both the two decomposed errors. Experimental results demonstrate that our method can find the optimal AAM subspace model rapidly and improve the performance of AAM significantly. Video Object SegmentationVideo Object Segmentation includes automatic segmentation and semi-automatic segmentation. For each of them, we proposed a statistical inference-based automatic segmentation method and a hierarchy optical flow based semi-automatic segmentation method respectively.- Backgroun...