A Probabilistic Cascade of Detectors for Individual Object Recognition

Abstract

Abstract.A probabilistic system for recognition of individual objects is presented. The objects to recognize are composed of constellations of features, and features from a same object share the common reference frame of the image in which they are detected. Features appearance and pose are modeled by probabilistic distributions, the parameters of which are shared across features in order to allow training from few examples.In order to avoid an expensive combinatorial search, our recognition system is organized as a cascade of well-established, simple and inexpensive detectors. The candidate hypotheses output by our algorithm are evaluated by a generative probabilistic model that takes into account each stage of the matching process.We apply our ideas to the problem of individual object recognition and test our method on several data-sets. We compare with Lowe's algorithm [7] and demonstrate significantly better performance.

Context, Computation, and Optimal ROC Performance in Hierarchical Models

“…Computation might be organized as a tree, for example to search simultaneously for multiple objects or to postpone decisions about pose or identity by exploring multiple branches, or as a cascade, which might be suitable for single objects and limited pose variation. The computational advantages are well documented, both from a practical and a theoretical standpoint (Fleuret and Geman 2001;Viola and Jones 2001;Moreels and Perona 2008;Blanchard and Geman 2005;Amit and Trouvé 2010).…”

Robust Homography Estimation from Local Affine Maps

“…This problem is recognized as difficult, especially under severe viewpoint changes between images. This is a fundamental step in many computer vision and image processing applications such as scene recognition [73,7,66,12,44,65,74,21,79,45] and detection [18,51], object tracking [81], robot localization [67,72,48,5,52], image stitching [2,6], image registration [78,30] and retrieval [22,20], 3D modeling and reconstruction [15,19,75,1,57], motion estimation [76], photo management [68,77,27,9], symmetry detection [32] or even image forgeries detection [10].…”