Ensemble-Based Tracking: Aggregating Crowdsourced Structured Time Series Data
Naiyan Wang, Dit‐Yan Yeung · Rare & Special e-Zone (The Hong Kong University of Science and Technology) · 2014
We study the problem of aggregating the con-tributions of multiple contributors in a crowd-sourcing setting. The data involved is in a form not typically considered in most crowdsourcing tasks, in that the data is structured and has a temporal dimension. In particular, we study the visual tracking problem in which the unknown data to be estimated is in the form of a sequence of bounding boxes representing the trajectory of the target object being tracked. We propose a factorial hidden Markov model (FHMM) for ensemble-based tracking by learning jointly the unknown trajectory of the target and the relia-bility of each tracker in the ensemble. For ef-ficient online inference of the FHMM, we de-vise a conditional particle filter algorithm by ex-ploiting the structure of the joint posterior dis-tribution of the hidden variables. Using the largest open benchmark for visual tracking, we empirically compare two ensemble methods con-structed from five state-of-the-art trackers with the individual trackers. The promising experi-mental results provide empirical evidence for our ensemble approach to “get the best of all worlds”. 1.