Real-time time-consistent 2D-to-3D video conversion based on color histogram

Abstract

In this paper we propose an algorithm for real-time 2D-to-3D conversion. Saliency based on color histogram is used to build depth map. To eliminate instability problem of color saliency, we presented a low-complexity and low-latency scene detection method taking advantage of color histogram produced before. The experiment results show that our algorithm is able to eliminate depth flickering effectively and create depth maps of good equality in various scenes in real-time.