Blur-Aware Disparity Estimation from Defocus Stereo Images

Defocus blur usually causes performance degradation in establishing the visual correspondence between stereo images. We propose a blur-aware disparity estimation method that is robust to the mismatch of focus in stereo images. The relative blur resulting from the mismatch of focus between stereo images is approximated as the difference of the square diameters of the blur kernels. Based on the defocus and stereo model, we propose the relative blur versus disparity (RBD) model that characterizes the relative blur as a second-order polynomial function of disparity. Our method alternates between RBD model update and disparity update in each iteration. The RBD model in return refines the disparity estimation by updating the matching cost and aggregation weight to compensate the mismatch of focus. Experiments using both synthesized and real datasets demonstrate the effectiveness of our proposed algorithm.

1. Introduction

Recovery of the depth information from binocular vision is an essential task since the depth information is useful in the reconstruction of 3D shape [12], matting [24], [14], and generating variable focus images/videos [13], [22]. However, the performance of stereo correspondence degrades in regions without prominent textures. Even with the support of the local neighborhood and regularization [2], [27], stereo matching remains a challenging problem.