Adaptive disparity estimation for stereoscopic video coding

Disparity estimation is an essential step in stereoscopic video coding. Existing disparity estimation methods are based on single disparity model and lack the robustness for estimating the real-time view disparities. To overcome this problem, this paper proposes an adaptive disparity estimation scheme which employs adaptive selection of disparity models based on the rate distortion optimization function. By exploiting the information obtained from the disparity models, the proposed technique estimates and compensates the disparities among different views in stereoscopic video coding efficiently. Experimental results show that the proposed technique contributes to improve the accurate of disparity estimation and save the bits spent on the coding.