RESUMO
A profile preferentially partial occlusion removal method for integral imaging is presented. The profile of the occlusion always contains details with significant texture structure, and regions with significant texture structure often lead to reliable depth estimation. Taking the advantage of the significant texture structure, the profile of occlusion is preferentially dealt with, and then the entire occlusion region is determined via regional spreading according to the accurate profile. The details of occlusion can be accurately removed and the occluded scene is also retained to the maximum degree. In our method, elemental images are integrated into a four-dimensional light field to provide consistently reliable depth estimation and occlusion decisions among all elemental images. Experimental results show that the proposed method is efficient to deal with the details of the occlusion, and it is robust for the occlusions with different kinds of texture structure.
