MTS: A Multiple Temporal Scale Tracker Handling Occlusion and Abrupt Motion Variation
نویسندگان
چکیده
We propose visual tracking over multiple temporal scales to handle occlusion and non-constant target motion. This is achieved by learning motion models from the target history at different temporal scales and applying those over multiple temporal scales in the future. These motion models are learned online in a computationally inexpensive manner. Reliable recovery of tracking after occlusions is achieved by extending the bootstrap particle filter to propagate particles at multiple temporal scales, possibly many frames ahead, guided by these motion models. In terms of the Bayesian tracking, the prior distribution at the current time-step is approximated by a mixture of the most probable modes of several previous posteriors propagated using their respective motion models. This improved and rich prior distribution, formed by the models learned and applied over multiple temporal scales, further makes the proposed method robust to complex target motion through covering relatively large search space with reduced sampling effort. Extensive experiments have been carried out on both publicly available benchmarks and new video sequences. Results reveal that the proposed method successfully handles occlusions and a variety of rapid changes in target motion.
منابع مشابه
Visual tracking over multiple temporal scales
Faculty of Science School of Computer Science Doctor of Philosophy by Muhammad Haris Khan Visual tracking is the task of repeatedly inferring the state (position, motion, etc) of the desired target in an image sequence. It is an important scientific problem as humans can visually track targets in a broad range of settings. However, visual tracking algorithms struggle to robustly follow a target...
متن کاملA Multiple Motion Model Tracker Handling Occlusion and Rapid Motion Variation
We propose a new tracking method capable of handling occlusions and non-constant target motion. This is achieved using multiple simple motion models, learned at different temporal scales and combined to capture possibly complex motion patterns. These motion models are learned online in a computationally inexpensive manner. Reliable recovery of tracking after occlusions is achieved by extending ...
متن کاملEffective Visual Tracking Using Multi-Block and Scale Space Based on Kernelized Correlation Filters
Accurate scale estimation and occlusion handling is a challenging problem in visual tracking. Recently, correlation filter-based trackers have shown impressive results in terms of accuracy, robustness, and speed. However, the model is not robust to scale variation and occlusion. In this paper, we address the problems associated with scale variation and occlusion by employing a scale space filte...
متن کاملFusion of Multiple Cues from Color and Depth Domains using Occlusion Aware Bayesian Tracker
Occlusion Aware Bayesian Tracker Kourosh MESHGI Shin-ichi MAEDA Shigeyuki OBA and Shin ISHII †Graduate School of Informatics, Kyoto University, Gokasho, Uji-shi, Kyoto, 611-0011 Japan E-mail: †{meshgi-k,ichi,oba,ishii}@sys.i.kyoto-u.ac.jp Abstract Object tracking has attracted considerable attention recently because of high demands for its everyday-life applications. Appearance-based trackers h...
متن کاملMulti - agent tracking under occlusion and 3 D motion interpretation
Tracking multiple objects in surveillance scenarios face considerable difficulty in handling occlusions. We report a composite tracker based on feature tracking and colour based tracking that demonstrates superior performance under high degrees of occlusion. Disjoint foreground blobs are extracted by using change masks obtained by combining an online-updated background model and flow informatio...
متن کامل