Method for detecting occlusions in an image, corresponding device and computer program product

Abstract

A method is proposed for detecting occlusions in an image captured by a light field capturing system, comprising, for at least one reference image belonging to a matrix of images captured by the light field capturing system: —calculating a depth map and a visibility map for pixels in the reference image; —determining at least one candidate area in the reference image in which the potential occlusion may occur based at least on a segmentation of the depth map; —determining an information representative of an occlusion state in the at least one candidate area based at least on visibility values of the visibility map associated with at least two neighborhoods of the at least one candidate area.

Dependencies are inferred from the claim text.