Solving Uncalibrated Photometric Stereo using Total Variation

Abstract

Estimating the shape and appearance of an object, given one or several images, is still an open and challenging research problem called 3D-reconstruction. Among the different techniques available, photometric stereo (PS) produces highly accurate results when the lighting conditions have been identified. When these conditions are unknown, the problem becomes the so-called uncalibrated PS problem, which is ill-posed. In this paper, we will show how total variation can be used to reduce the ambiguities of uncalibrated PS, and we will study two methods for estimating the parameters of the generalized bas-relief ambiguity. These methods will be evaluated through the 3D-reconstruction of real-world objects.