A Novel Shape-from-shading Algorithm using Jacobi Iterative Method and Bi-directional Estimation

O. Ikeda (Japan)

Keywords

shape, shapefromshading, Jacobi’s iterative method

Abstract

Recent research on shape-from-shading has been made with focus on improvements of the minimization approaches; several authors have presented sophisticated constraints and algorithms to solve the problem of over-smoothing. A linear approach has also attracted attention due to its simple implementation. We present in this paper a new nonlinear approach, which is as simple in implementation as the linear approach and as accurate as those minimization ones. In our method, the image irradiance equation is represented by the depth parameter and the Jacobi iterative method is applied, to get a matrix form of iterative relation. Two shapes are estimated in the forward and backward directions and are averaged to minimize shape distortions. Some convincing numerical results are given.

Important Links:



Go Back