Wang-Landau Monte Carlo-Based Tracking Methods for Abrupt Motions

Abstract

We propose a novel tracking algorithm based on the Wang-Landau Monte Carlo (WLMC) sampling method for dealing with abrupt motions efficiently. Abrupt motions cause conventional tracking methods to fail because they violate the motion smoothness constraint. To address this problem, we introduce the Wang-Landau sampling method and integrate it into a Markov Chain Monte Carlo (MCMC)-based tracking framework. By employing the novel density-of-states term estimated by the Wang-Landau sampling method into the acceptance ratio of MCMC, our WLMC-based tracking method alleviates the motion smoothness constraint and robustly tracks the abrupt motions. Meanwhile, the marginal likelihood term of the acceptance ratio preserves the accuracy in tracking smooth motions. The method is then extended to obtain good performance in terms of scalability, even on a high-dimensional state space. Hence, it covers drastic changes in not only position but also scale of a target. To achieve this, we modify our method by combining it with the N-fold way algorithm and present the N-Fold Wang-Landau (NFWL)-based tracking method. The N-fold way algorithm helps estimate the density-of-states with a smaller number of samples. Experimental results demonstrate that our approach efficiently samples the states of the target, even in a whole state space, without loss of time, and tracks the target accurately and robustly when position and scale are changing severely.

Abrupt motion tracking via nearest neighbor field driven stochastic sampling

“…Kwon et al [17] propose to utilize the Wang-Landau Monte Carlo(WLMC) sampling method to deal with the local-trap problem in abrupt motion tracking. Along with this thread, in…”

“…Compared with [17], the posterior distribution can be more effectively estimated by a stochastic approximation process. However, this method has to explore the whole sample space uniformly with an inefficient preliminary sampling phase.…”

“…1. We compare the proposed method with other 6 state-of-the-art algorithms: WLMC [17], SAMC [18], SCM [9], VTD [38], LSST [7], and saliencybased particle filter (referred as SaPF) [24]. For fair comparison, we run the source codes provided by the authors with tuned parameters to obtain their best performance.…”

“…In abrupt motion tracking, it is commonly assumed that the target almost moves smoothly but abruptly changes over several frames. However, most trackers [17,18] simply consider all moves as abrupt changes and thus tend to suffer from drifting in case of background clutter or distractions.…”

Adaptive fusion of particle filtering and spatio-temporal motion energy for human tracking

“…On the other hand, the Wang-Landau sampling methods was integrated with a Markov Chain Monte Carlo (MCMC) based tracking framework. This was achieved by applying the density-of-states term estimated by the Wang-Landau sampling method to the acceptance ratio of MCMC to alleviate the motion smoothness constraint [46]. Schaap et al [47] introduced a Bayesian tube tracking algorithm that incorporated a priori knowledge to enhance the tracking performance of tube structures.…”