A robust color correction method for stereoscopic video coding

Color inconsistency is an urgent problem to be solved in stereoscopic video coding. In this paper, a robust color correction method for stereoscopic video coding is proposed. We first perform disparity estimation between right image and left image, and detect static scene between the current right image and the corrected right images, respectively. Next, spatial and temporal color correction matrices are estimated for right view by using linear regression model, and are combined to get the final corrected view. Finally, the corrected view is coded based on the prediction structure. Experimental results show that the proposed method can produce well correction quality for multiview video. Additionally, higher coding efficiency can be achieved for multi-view video.