Adaptive appearance learning for visual object tracking

Object Tracking, 2011

Image and Vision Computing, 2007

Visual tracking has been a challenging problem in computer vision over the decades. The applications of visual tracking are far-reaching, ranging from surveillance and monitoring to smart rooms. Mean-shift tracker, which gained attention recently, is known for tracking objects in a cluttered environment. In this work, we propose a new method to track objects by combining two well-known trackers, sum-of-squared differences (SSD) and color-based mean-shift (MS) tracker. In the proposed combination, the two trackers complement each other by overcoming their respective disadvantages. The rapid model change in SSD tracker is overcome by the MS tracker module, while the inability of MS tracker to handle large displacements is circumvented by the SSD module. The performance of the combined tracker is illustrated to be better than those of the individual trackers, for tracking fast-moving objects. Since the MS tracker relies on global object parameters such as color, the performance of the tracker degrades when the object undergoes partial occlusion. To avoid adverse effects of the global model, we use MS tracker to track local object properties instead of the global ones. Further, likelihood ratio weighting is used for the SSD tracker to avoid drift during partial occlusion and to update the MS tracking modules. The proposed tracker outperforms the traditional MS tracker as illustrated.

Journal of Signal Processing Systems

This paper addresses the issue of tracking a single visual object through crowded scenarios, where a target object may be intersected or partially occluded by other objects for a long duration, experience severe deformation and pose changes, and different motion speed in cluttered background. A robust visual object tracking scheme is proposed that exploits the dynamics of object shape and appearance similarity. The method uses a particle filter where a multi-mode anisotropic mean shift is embedded to improve the initial particles. Comparing with the conventional particle filter and mean shift-based tracking (Shan et al. 2004), our method offers the following novelties: We employ a fully tunable rectangular bounding box described by five parameters (2D central location, width, height, and orientation) and full functionaries in the joint tracking scheme; We derive the equations for the multi-mode version of the anisotropic mean shift where the rectangular bounding box is partitioned into concentric areas, allowing better tracking objects with multiple modes. The bounding box parameters are then computed by using eigen-decomposition of mean shift estimates and weighted averaging. This enables a more efficient re-distributions of initial particles towards locations associated with large weights, hence an efficient particle filter tracking using a very small number of particles (N = 15 is used). Experiments have been conducted on video containing a range of complex scenarios, where tracking results are further evaluated by using two objective criteria and compared with two existing tracking methods. Our results have shown that the propose method is robust in terms of tracking drift, tightness and accuracy of tracked bounding boxes, especially in scenarios where the target object contains long-term partial occlusions, intersections, severe deformation, pose changes, or cluttered background with similar color distributions.

2010 20th International Conference on Pattern Recognition, 2010

This paper proposes a novel visual object tracking scheme, exploiting both local point feature correspondences and global object appearance using the anisotropic mean shift tracker. Using a RANSAC cost function incorporating the mean shift motion estimate, motion smoothness and complexity terms, an optimal feature point set for motion estimation is found even when a high proportion of outliers is presented. The tracker dynamically maintains sets of both foreground and background features, the latter providing information on object occlusions. The mean shift motion estimate is further used to guide the inclusion of new point features in the object model. Our experiments on videos containing long term partial occlusions, object intersections and cluttered or close color distributed background have shown more stable and robust tracking performance in comparison to three existing methods.

IEEE Transactions on Pattern Analysis and Machine Intelligence, 2000

We propose a similarity measure based on a Spatial-color Mixture of Gaussians (SMOG) appearance model for particle filters. This improves on the popular similarity measure based on color histograms because it considers not only the colors in a region but also the spatial layout of the colors. Hence, the SMOG-based similarity measure is more discriminative. To efficiently compute the parameters for SMOG, we propose a new technique, with which the computational time is greatly reduced. We also extend our method by integrating multiple cues to increase the reliability and robustness. Experiments show that our method can successfully track objects in many difficult situations.

2007

This paper presents a robust object tracking method via a spatial bias appearance model learned dynamically in video. Motivated by the attention shifting among local regions of a human vision system during object tracking, we propose to partition an object into regions with different confidences and track the object using a dynamic spatial bias appearance model (DSBAM) estimated from region confidences. The confidence of a region is estimated to reflect the discriminative power of the region in a feature space, and the probability of occlusion. We propose a novel hierarchical Monte Carlo (HAMC) algorithm to learn region confidences dynamically in every frame. The algorithm consists of two levels of Monte Carlo processes implemented using two particle filtering procedures at each level and can efficiently extract high confidence regions through video frames by exploiting the temporal consistency of region confidences. A dynamic spatial bias map is then generated from the high confidence regions, and is employed to adapt the appearance model of the object and to guide a tracking algorithm in searching for correspondences in adjacent frames of video images. We demonstrate feasibility of the proposed method in video surveillance applications. The proposed method can be combined with many other existing tracking systems to enhance the robustness of these systems.

BMVC, 2002

Procedia Computer Science, 2017

Even though many algorithms have been developed and many applications of object tracking have been made, object tracking is still considered as a difficult task to accomplish. The existence of several problems such as illumination variation, tracking non-rigid object, non-linear motion, occlusion, and requirement of real time implementation has made tracking as one of the challenging tasks in computer vision. In this paper a tracking algorithm which combines mean-shift and particle-Kalman filter is proposed to overcome above mentioned problems. The purpose of this combination is to draw each algorithm's strength points and cover each algorithms drawbacks. In the proposed method, mean-shift is used as master tracker when the target object is not occluded. When occlusion is occurred or the mean-shift tracking result is not convincing, particle-Kalman filter will act as master tracker to improve the tracking results. Experimental results of the proposed method show desirable performance in tracking objects under several above mentioned problems.

2009 IEEE International Conference on Multimedia and Expo, 2009

We propose a novel scheme that jointly employs anisotropic mean shift and particle filters for tracking moving objects from video. The proposed anisotropic mean shift, that is applied to partitioned areas in a candidate object bounding box whose parameters (center, width, height and orientation) are adjusted during the mean shift iterations, seeks multiple local modes in spatial-kernel weighted color histograms. By using a Gaussian distributed Bhattacharyya distance as the likelihood and mean shift updated parameters as the state vector, particle filters become more efficient in terms of tracking using a small number of particles (<20). The combined scheme is able to maintain the merits of both methods. Experiments conducted on videos containing deformable objects with long-term partial occlusions and intersections have shown robust tracking performance. Comparisons with two existing methods have been made which showed marked improvement in terms of robustness to occlusions, tightness and accuracy of tracked box, and tracking drift.