Issue No. 03 - March (2010 vol. 32)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/TPAMI.2009.23
Mustafa Özuysal , Ecole Polytechnique Fédérale de Lausanne, Lausanne
Michael Calonder , Ecole Polytechnique Fédérale de Lausanne, Lausanne
Vincent Lepetit , Ecole Polytechnique Fédérale de Lausanne, Lausanne
Pascal Fua , Ecole Polytechnique Fédérale de Lausanne, Lausanne
While feature point recognition is a key component of modern approaches to object detection, existing approaches require computationally expensive patch preprocessing to handle perspective distortion. In this paper, we show that formulating the problem in a naive Bayesian classification framework makes such preprocessing unnecessary and produces an algorithm that is simple, efficient, and robust. Furthermore, it scales well as the number of classes grows. To recognize the patches surrounding keypoints, our classifier uses hundreds of simple binary features and models class posterior probabilities. We make the problem computationally tractable by assuming independence between arbitrary sets of features. Even though this is not strictly true, we demonstrate that our classifier nevertheless performs remarkably well on image data sets containing very significant perspective changes.
Image processing and computer vision, object recognition, tracking, image registration, feature matching, naive Bayesian.
M. Özuysal, P. Fua, V. Lepetit and M. Calonder, "Fast Keypoint Recognition Using Random Ferns," in IEEE Transactions on Pattern Analysis & Machine Intelligence, vol. 32, no. , pp. 448-461, 2009.