Using a low pass filter to recover three-dimensional shape from focus in the presence of noise

Abstract

Several passive methods like Shape from Focus (SFF) have been proposed for recovering 3-D shape of the objects from their 2-D images. Presence of noise in the images affects the accuracy of shape recovery. Most of the existing approaches show feeble performance when noise is present in the images. In this paper, we propose a new method based On a simple low pass filter, especially designed for SFF methods. The proposed scheme is experimented and evaluated using different image sequences of synthetic and real objects. It provides better results as compared with previous approaches; especially its performance is impressive for noisy images.