Authors : Israa Hadi Ali and M. Mustafa Ehsan
Abstract: In this study, proposes an efficient method for video inpainting that based on statistical modified background modeling. Deferent challenges and situations can be handled in this algorithm with a good result and less computational cost. There are three major steps in this proposed methodology, in the first step the proposed system makes an estimation to the background of the video to get the background. The second step which is subtracted each frame of input video from the background model in order to get the foreground objects to be deleted. Third step the method detects the trajectory of the target object that selected by the user and shutting off its pixels. Finally, finding the appropriate pixels from the background to be assigned instead of the target object pixels.
Israa Hadi Ali and M. Mustafa Ehsan, 2018. Video Inpainting Based on Background Modelling. Journal of Engineering and Applied Sciences, 13: 1382-1385.