An iterative marching with correctness criterion algorithm for shape from shading under oblique light source

Abstract

In this paper, a fast and robust Shape from Shading (SfS) algorithm by iterative marching with corrections criterion under oblique light source is presented. Usually, SfS algorithms are based on the assumption that image radiance is a function of normal surface alone. SfS algorithms solve first-order nonlinear Hamilton Jacobi equation called image irradiance equation. Both Fast Marching Method (FMM) and Marching with Correctness Criterion (MCC) basically work for the frontal light illumination direction, in which the image irradiance equation is an Eikonal equation. The problem task is to recover the surface from the image—which amounts to finding a solution to the Eikonal equation.FMMcopes better the image irradiance iteratively under oblique light sources with the cost of computational complexity O(NlogN). One prominent solution is the Marching with MCC of Mauch which solves the Eikonal equation with computational complexity O(N). Here, we present a new iterative variant of the MCC which copes better with images taken under oblique light sources. The proposed approach is evaluated on two synthetic real images and compared with the iterative variant of FMM. The experimental results show that the proposed approach, iterative variant of MCC is more efficient than the iterative variant of FMM.