Depth Recovery of Complex Surfaces from Texture-less Pairs of Stereo Images

In this paper, a novel framework is presented to recover the 3D shape information of a complex surface using its texture-less stereo images. First a linear and generalized Lambertian model is proposed to obtain the depth information by shape from shading (SfS) using an image from stereo pair. Then this depth data is corrected by integrating scale invariant features (SIFT) indexes. These SIFT indexes are defined by means of disparity between the matching invariant features in rectified stereo images. The integration process is based on correcting the 3D visible surfaces obtained from SfS using these SIFT indexes. The SIFT indexes based improvement of depth values which are obtained from generalized Lambertian reflectance model is performed by a feed-forward neural network. The experiments are performed to demonstrate the usability and accuracy of the proposed framework.