Stereoscopic Video Color Correction Based on Keyframe Extraction

Color inconsistency is an urgent problem to be solved in stereoscopic video. In this paper, a robust color correction method for stereoscopic video is proposed. We first take advantage of feature points matching based on Scale Invariant Feature (SIFT) to find the corresponding match points of the images. Next, parameter model between the images is calculated by the value of each match point pair and obtained the optimal from the least square method. Finally, key frame extraction technology based on color feature is used to perform consistently the color correction of parameter model for each key frame and its next every frame. Experimental results show that the proposed method can produce well correction quality for stereoscopic video.