Occlusion-aware depth estimation for light field using multi-orientation EPIs

摘要

Highlights?We introduced a method to extract various directions of EPIs and showed that the optimal EPI has the same direction as the occlusion boundary.?We proposed to detect the occlusion boundaries based on the variance and gradient of local depth values, and find the accurate depth label from the optimal EPI.?We developed a novel framework that combines the local disparity map with the occlusion boundary to deal with occlusions so as to obtain robust depth maps.AbstractEpipolar plane images (EPIs) contain special linear structures that reflect the disparity of a 3D point and are widely used in light field depth estimation. However, previous EPI-based approaches only utilize horizontal and vertical EPIs to estimate local disparities and ignore diagonal directions. In order to make full use of the regular grid light field images, we develop a strategy to extract epipolar plane images in all available directions. Based on the multi-orientation EPIs, a specific EPI in which