Depth from Shading Based on 2D Maximum Entropy

Depth from shading is an open problem in computer vision. During the estimation of object depth or shape, integrability constraint is enforced to regularize the solution. In this paper, we introduced a regularizing term of smoothness constraint employing the 2D maximum entropy to enforce the integrability. The test on synthetic image shows the proposed method has the better performance than the available methods. It is also robust to the noise, so there is a good application foreground for the method.