Separating Transparent Layers and Photometric Modeling

点击翻译按钮获取中文摘要

Looking through a window, the object behind the window is often disturbed by a reflection of another object One of the necessary techniques for constructing virtual museum is the technique to measure the artwork arranged inside a glass/acrylic display case without bringing the artwork out from the display case. To solve these problems, we developed a method of separating mixtures of two images, a method of separating mixtures of an arbitrary number of images, and a photometric stereo method. The first method of separating mixtures of two images is realized by classifying edges in the input image based on the belonging layer and formalizing the problem of decomposing the single image into two layer images as an optimization problem easier to solve based on this classification, and then solving this optimization with a deterministic annealing. The second method of separating mixtures of an arbitrary number of images accomplishes the blind separation of an arbitrary number of images from a set of their linear mixtures with unknown coefficients. This method first introduces a novel multiple correlation between one image and a set of multiple images. Then this multiple correlation leads us to provide a set of simultaneous linear equations for updating each mixture of images. Finally, source images are recovered by iterating between solving the sets of equations and cyclically permuting the mixtures of images. The third photometric stereo method estimates the shape and the reflectance of the artwork arranged inside a transparent display case. If we illuminate the display case, the light will reflect at its surface, thus, we cannot apply conventional photometric stereo as is. This method adopts a five-light photometric stereo that estimates the shape and the reflectance of the object which has specularities, under the circumstances that the light is reflected at the display case.