A Probabilistic Correspondence Algorithm Using Shape Cues and Prior Information

Correspondence of features between two or more images obtained from different views of the same object is still a challenging problem in vision. In many applications, these two views are part of a larger video sequence. The aim of this paper is to show that the original video data contains enough information (referred to as prior information) to design a robust correspondence algorithm. A doubly stochastic matrix, representing the probability of match between the features, is derived using the Sinkhorn normalization procedure. The final correspondence is obtained by minimizing the probability of error of a match between the entire constellation of features in the two sets, thus taking into account the spatial structure of the object. The method is applied for creating holistic 3D models from partial ones and the results are presented.