A second-order theory of texture for depth from focus—interactive supplement

Each scene visualizes reconstructions using both depth map images and surface renders of the associated point sets.

Reconstructions marked "fixed" use the fixed step-size DFF algorithm (Algorithm 1 in the PDF supplement). Reconstructions marked "adaptive" use the pixel-adaptive DFF algorithm (Algorithm 2 in the PDF supplement). Reconstructions marked "progressive" use a GIF animation to show the impact of progressively increasing effective patch size through additional aggregations.

Some scenes show a photo from an alternative color camera (not used for reconstruction), to provide better context for the scene.