Shape and Reflectance from Image Intensities

This paper presents new results in the area of image intensity understanding. We address the problem of recovering three-dimensional shapes of surfaces from twc-dimensional intensity images. Previous shape-fromintensity methods are not applicable to surfaces with varying reflectance properties. These methods are also inaccurate in the case of concave surfaces as they do not account for interreflections. Here, we present shape recovery methods that overcome both these problems. Our solutions are based on physics-based models of reflection and interreflection. Multiple images of a surface are obtained by varying the source direction. Shape and reflectance parameters are computed simultaneously from the measured image intensities. Experimental results are included to demonstrate the accuracy and robustness of the proposed shape recovery methods.