Sparkle Vision : Seeing the World through Random

This thesis studies the problem of reproducing the world lighting from a single image of an object covered with random specular microfacets on the surface. Such a reflector can be interpreted as a randomized mapping from the lighting to the image. This intrinsic randomness makes it challenging for humans to interpret the image of a specular surface. We propose a system to solve it algorithmically and demonstrate how a simple yet reliable method can calibrate the proposed system and do the inference. The success of such system relies on accurate exposure of the specular surfaces. However, such objects have very distinguished optical properties compared with both diffuse surfaces and smooth specular objects like metals. So we design a special imaging system to robustly and effectively photograph them. Finally we conduct experiments to verify the correctness of our model assumptions and prove the effectiveness of our pipeline. Thesis Supervisor: Edward H. Adelson Title: John and Dorothy Wilson Professor of Vision Science