Study of Shape from Focus Algorithm

Abstract

Computer vision incorporates various three dimensional (3D) vision techniques which mainly aims at reconstructing a scene from a set of two dimensional (2D) images. Passive 3D vision techniques such as stereo vision require disparity between the images of the same scene taken from multiple views. Shape from focus (SFF) is a technique which recovers 3D geometry of the scene based on focus which is an accommodation cue. Multiple images of the scene are taken using a single camera with varying focus levels thus avoiding complex and expensive vision systems for reconstruction. However, one of the main drawbacks of this technique is that it requires the imaged scene to have texture variation for effective reconstruction. In this paper a SFF system has been developed and working of SFF algorithm for synthetic images has been studied.