Single and Multiple Object Tracking Using Log-Euclidean Riemannian Subspace and Block-Division Appearance Model
Issue No. 12 - Dec. (2012 vol. 34)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/TPAMI.2012.42
Weiming Hu , Nat. Lab. of Pattern Recognition, Inst. of Autom., Beijing, China
Xi Li , Nat. Lab. of Pattern Recognition, Inst. of Autom., Beijing, China
Wenhan Luo , Nat. Lab. of Pattern Recognition, Inst. of Autom., Beijing, China
Xiaoqin Zhang , Nat. Lab. of Pattern Recognition, Inst. of Autom., Beijing, China
S. Maybank , Dept. of Comput. Sci. & Inf. Syst., Birkbeck Coll., London, UK
Zhongfei Zhang , Dept. of Comput. Sci., Binghamton Univ., Binghamton, NY, USA
Object appearance modeling is crucial for tracking objects, especially in videos captured by nonstationary cameras and for reasoning about occlusions between multiple moving objects. Based on the log-euclidean Riemannian metric on symmetric positive definite matrices, we propose an incremental log-euclidean Riemannian subspace learning algorithm in which covariance matrices of image features are mapped into a vector space with the log-euclidean Riemannian metric. Based on the subspace learning algorithm, we develop a log-euclidean block-division appearance model which captures both the global and local spatial layout information about object appearances. Single object tracking and multi-object tracking with occlusion reasoning are then achieved by particle filtering-based Bayesian state inference. During tracking, incremental updating of the log-euclidean block-division appearance model captures changes in object appearance. For multi-object tracking, the appearance models of the objects can be updated even in the presence of occlusions. Experimental results demonstrate that the proposed tracking algorithm obtains more accurate results than six state-of-the-art tracking algorithms.
video signal processing, Bayes methods, cameras, covariance matrices, feature extraction, inference mechanisms, learning (artificial intelligence), object tracking, particle filtering (numerical methods), particle filtering-based Bayesian state inference, single object tracking, multiple object tracking, incremental log-Euclidean Riemannian subspace learning algorithm, block-division appearance model, object appearance modeling, videos, nonstationary cameras, occlusion reasoning, symmetric positive definite matrices, covariance matrices, image features, vector space, log-Euclidean Riemannian metric, log-Euclidean block-division appearance model, global spatial layout information, local spatial layout information, Tracking, Cameras, Solid modeling, Covariance matrix, Algorithm design and analysis, Inference algorithms, Visual analytics, block-division appearance model, Visual object tracking, occlusion reasoning, log-euclidean Riemannian subspace, incremental learning
Xiaoqin Zhang, Wenhan Luo, Xi Li, Weiming Hu, S. Maybank and Zhongfei Zhang, "Single and Multiple Object Tracking Using Log-Euclidean Riemannian Subspace and Block-Division Appearance Model," in IEEE Transactions on Pattern Analysis & Machine Intelligence, vol. 34, no. , pp. 2420-2440, 2012.