VIDEO COMPLETION USING HIERARCHICAL MOTION ESTIMATION AND COLOR COMPENSATION

Jin-Hong Kim1 , Rae-Hong Park1 and Jino Lee1 1Department of Electronic Engineering, Sogang University, Seoul, Korea  

ABSTRACT 

Video completion is used for repairing damaged or missing data in a video sequence. This paper proposes a video completion method using hierarchical motion estimation and color compensation. The proposed method consists of three steps: hierarchical motion estimation, source patch search, and color assignment. If an input video contains brightness change, existing video completion methods give poor results. For overcoming this drawback, we apply two methods: color normalization for motion estimation and source patch search, and color compensation for color assignment. First of all, we construct spatio-temporal Gaussian pyramid for reducing the computation time. We estimate the motion of the input video using color information of each pyramid level. The proposed method uses normalized RGB color in motion estimation. Then, we search the source patch using normalized color information and motion information extracted in the first step. Finally, color compensation using color transform is used for color assignment. Experimental results show that the proposed method gives better visual results than conventional video completion methods. 

KEYWORDS 

Video completion, motion estimation, Gaussian pyramid

More Details : http://airccse.org/journal/ijcga/papers/3113ijcga02.pdf