Off-line multiple object tracking using candidate selection and the Viterbi algorithm

Abstract

This paper presents a probabilistic framework for off-line multiple object tracking. At each timestep, a small set of deterministic candidates is generated which is guaranteed to contain the correct solution. Tracking an object within video then becomes possible using the Viterbi algorithm. In contrast with particle filter methods where candidates are numerous and random, the proposed algorithm involves a few candidates and results in a deterministic solution. Moreover, we consider here off-line applications where past and future information is exploited. This paper shows that, although basic and very simple, this candidate selection allows the solution of many tracking problems in different real-world applications and offers a good alternative to particle filter methods for off-line applications.

Closed-world tracking of multiple interacting targets for indoor-sports applications

“…The class of visual representations that can be viewed as a generalization of this approach is the color histograms [52]. They have been successfully applied in many tracking applications in sports [39,12,43] as well as in the more general applications of visual tracking [40,37,54,13]. Recently, Birchfield and Rangarajan [7] proposed a class of color histograms that also integrates the spatial information of the target's color.…”

“…The problem of tracking is then posed as the task of finding the optimal path through the graph [19,43,51]. These approaches, however, rely on the explicit detection of all the players on the court using background-subtraction techniques.…”

Region of interest extraction using colour based methods on the CUAVE Database

“…None of the methods failed consistently for any one individual. Hence by employing a Viterbi tracking algorithm [13] it would be possible to prevent isolated errors causing an AVSR system to lose track of the ROI.…”

A Tracking Framework for Accurate Face Localization

“…The frames, where face detection took place can have states D, F and B, while the other frames can have states F and B only. The complexity of the trellis is considerably reduced in comparison with other approaches that draw the trellis using all the bounding boxes provided by the detector or the tracker [6]. In fact, the number of possible paths in the trellis grows exponentially with the number of nodes.…”

“…They were also used to perform the face detection and tracking, searching for the best matching region for a given face template [5]. In [6], a multiple object tracking is presented, where the Viterbi Algorithm is used to find the best path between candidates selected according to skin color criteria. In this paper, a new deterministic approach is presented.…”