Abstract
This paper presents a novel and general method for thedetection, rectification and segmentation of imaged copla-nar repeated patterns. The only assumption made of thescene geometry is that repeated scene elements are mappedto each other by planar Euclidean transformations. Theclass of patterns covered is broad and includes nearly allcommonly seen, planar, man-made repeated patterns. Inaddition, novel linear constraints are used to reduce geo-metric ambiguity between the rectified imaged pattern andthe scene pattern. Rectification to within a similarity of thescene plane is achieved from one rotated repeat, or to withina similarity with a scale ambiguity along the axis of symme-try from one reflected repeat. A stratum of constraints is derived that gives the necessary configuration of repeats for each successive level of rectification. A generative model for the imaged pattern is inferred and used to segment the pattern with pixel accuracy. Qualitative results are shown on a broad range of image types on which state-of-the-artmethods fail.