Depth Masks for Image Segmentation for Depth-based Computational Photography

摘要: A depth-based computational photography method includes recording a first image using stereoscopic camera system, establishing disparity map, via processor, of the including determination pixel distance between features in and second image, determining histogram values from removing statistical outlier values. The further identifying holes map computationally shrinking averaging based on radially sampled areas proximate to hole sharpening value peaks by shifting close neighborhood range peak maximum closer at maximum.