Abstract:A novel visual object tracking scheme is proposed by using joint point feature correspondences and object appearance similarity. For point feature-based tracking, we propose a candidate tracker that simultaneously exploits two separate sets of point feature correspondences in the foreground and in the surrounding background, where background features are exploited for the indication of occlusions. Feature points in these two sets are then dynamically maintained. For object appearance-based tracking, we propose a candidate tracker based on an enhanced anisotropic mean shift with a fully tunable (five degrees of freedom) bounding box that is partially guided by the above feature point tracker. Both candidate trackers contain a reinitialization process to reset the tracker in order to prevent accumulated tracking error propagation in frames. In addition, a novel online learning method is introduced to the enhanced mean shift-based candidate tracker. The reference object distribution is updated in each time interval if there is an indication of stable and reliable tracking without background interferences. By dynamically updating the reference object model, tracking is further improved by using a more accurate object appearance similarity measure. An optimal selection criterion is applied to the final tracker based on the results of these candidate trackers. Experiments have been conducted on several videos containing a range of complex scenarios. To evaluate the performance, the proposed scheme is further evaluated using three objective criteria, and compared with two existing trackers. All our results have shown that the proposed scheme is very robust and has yielded a marked improvement in terms of tracking drift, tightness, and accuracy of tracked bounding boxes, especially for complex video scenarios containing long-term partial occlusions or intersections, deformation, or background clutter with similar color distributions to the foreground object.

Joint Registration and Active Contour Segmentation for Object Tracking

Robust Visual Tracking Via CAMShift and Structural Local Sparse Appearance Model

Automatic Video Object Segmentation and Tracking Using an Improved Active Contour

Joint Feature Correspondences and Appearance Similarity for Robust Visual Object Tracking

A fast region-based active contour for non-rigid object tracking and its shape retrieval

JOTS: Joint Online Tracking and Segmentation

Video Object Contour Tracking Using Improved Dual-Front Active Contour

An Active Contour Tracking Method By Matching Foreground And Background Simultaneously

Jpdaf Based Hmm for Real-Time Contour Tracking

JROTM: Jointly Reinforced Object Tracking with Temporal Content Reference and Motion Guidance

Joint Conditional Random Field Filter for Multi-Object Tracking

Real-Time Tracking Combined with Object Segmentation

Multiple Cues-Based Active Contours for Target Contour Tracking under Sophisticated Background

Non-Rigid Object Tracking As Salient Region Segmentation And Association

Structured Partial Least Squares for Simultaneous Object Tracking and Segmentation.

Jointly Modeling Motion and Appearance Cues for Robust RGB-T Tracking

Deep Active Contours for Real-time 6-Dof Object Tracking

Robust Object Tracking Using Joint Color-Texture Histogram

Active-matting-based Object Tracking with Color Cues

Joint Detection, Re-Identification, And Lstm In Multi-Object Tracking

Tri-tracking Method for Robust Object Tracking