DocumentCode :
55098
Title :
Joint Individual-Group Modeling for Tracking
Author :
Bazzani, Loris ; Zanotto, Matteo ; Cristani, Marco ; Murino, Vittorio
Author_Institution :
Pattern Anal. & Comput. Vision (PAVIS), Ist. Italiano di Tecnol., Genoa, Italy
Volume :
37
Issue :
4
fYear :
2015
fDate :
April 1 2015
Firstpage :
746
Lastpage :
759
Abstract :
We present a novel probabilistic framework that jointly models individuals and groups for tracking. Managing groups is challenging, primarily because of their nonlinear dynamics and complex layout which lead to repeated splitting and merging events. The proposed approach assumes a tight relation of mutual support between the modeling of individuals and groups, promoting the idea that groups are better modeled if individuals are considered and vice versa. This concept is translated in a mathematical model using a decentralized particle filtering framework which deals with a joint individual-group state space. The model factorizes the joint space into two dependent subspaces, where individuals and groups share the knowledge of the joint individual-group distribution. The assignment of people to the different groups (and thus group initialization, split and merge) is implemented by two alternative strategies: using classifiers trained beforehand on statistics of group configurations, and through online learning of a Dirichlet process mixture model, assuming that no training data is available before tracking. These strategies lead to two different methods that can be used on top of any person detector (simulated using the ground truth in our experiments). We provide convincing results on two recent challenging tracking benchmarks.
Keywords :
computer vision; image classification; learning (artificial intelligence); object tracking; particle filtering (numerical methods); probability; Dirichlet process mixture model; classifier training; decentralized particle filtering framework; joint individual-group distribution; joint individual-group modeling; joint individual-group state space; merging event; object tracking; online learning; person detector; probabilistic framework; splitting event; Analytical models; Approximation methods; Detectors; Joints; Mathematical model; Merging; Monte Carlo methods; Dirichlet process mixture model; Group modeling; decentralized particle filtering; joint individual-group tracking;
fLanguage :
English
Journal_Title :
Pattern Analysis and Machine Intelligence, IEEE Transactions on
Publisher :
ieee
ISSN :
0162-8828
Type :
jour
DOI :
10.1109/TPAMI.2014.2353641
Filename :
6891328
Link To Document :
بازگشت