A Probabilistic Framework for Multitarget Tracking with Mutual Occlusions

Abstract

Mutual occlusions among targets can cause track loss or target position deviation, because the observation likelihood of an occluded target may vanish even when we have the estimated location of the target. This paper presents a novel probability framework for multitarget tracking with mutual occlusions. The primary contribution of this work is the introduction of a vectorial occlusion variable as part of the solution. The occlusion variable describes occlusion states of the targets. This forms the basis of the proposed probability framework, with the following further contributions: 1) Likelihood: A new observation likelihood model is presented, in which the likelihood of an occluded target is computed by referring to both of the occluded and occluding targets. 2) Priori: Markov random field (MRF) is used to model the occlusion priori such that less likely "circular" or "cascading" types of occlusions have lower priori probabilities. Both the occlusion priori and the motion priori take into consideration the state of occlusion. 3) Optimization: A realtime RJMCMC-based algorithm with a new move type called "occlusion state update" is presented. Experimental results show that the proposed framework can handle occlusions well, even including long-duration full occlusions, which may cause tracking failures in the traditional methods.

UA-DETRAC: A new benchmark and protocol for multi-object detection and tracking

“…Numerous multi-object tracking methods formulated the task as the state estimation problem using the filter based strategies, such as Kalman filter [38,36] and particle filter [29,32,39,63]. These methods typically predicted the states of targets in short time durations which did not perform well in complex scenarios.…”

FollowMe: Efficient Online Min-Cost Flow Tracking with Bounded Memory and Computation

“…Discrete-continuous optimization techniques have been proposed [3,38], where discrete (data association) and continuous (trajectory fitting) optimization problems are solved in an alternating way until converging to a hopefully better local minimum. In [7,11,15,17,59], approximate Markov chain Monte Carlo techniques are employed for solving the data association problem. In order to increase the discriminative power of appearance and dynamical models, online learning approaches have been suggested [33, 35, 46, 56-58, 60, 62].…”

Learning Discriminative Hidden Structural Parts for Visual Tracking

“…The MH algorithm [19] has been applied in the multiple target tracking task [21,23], where the authors use it for particle filter sampling to identify the state of each target precisely. However, in our paper, we do not aim at identifying each part all the time.…”