Abstract:Visual object tracking is a challenging and essential research problem in the field of computer vision. Recent years, many Online Multiple Instance Learning (MIL) tracking methods have been proposed with promising experimental results. These methods train a discriminative classifier under the boosting framework. The weak classifiers are learned from parts of the object and all classifiers are updated while updating the appearance model. However, due to irregular shape of object or occlusions, some components are not on object and should not be learned. On the contrary, a discriminative weak classifier learned from these components will mislead the tracker to drift away. To overcome this problem, we propose a novel online MIL tracking approach by updating with reliable components (OMRC). It keeps both background and object templates while tracking. By comparing current tracking result with two templates, we can get whether the pixels belong to object. The components which have a higher rate of pixels belong to object than a predefined threshold are reliable components. Moreover, in order to represent images better, we use HOG features and Histogram features instead of the widely used Haar-like features and design a new online weak classifier learning method. Experiments are performed on two challenging datasets including OTB2015 and Temple Color. Experimental results demonstrate the robustness of our OMRC tracker and the effectiveness of each component in the OMRC tracker.

Object Tracking of Online Feature Selection Multiple Instance Learning Via Minimizing Trace

Mean Shift Tracking Based on Ensemble Multiple Instance Learning

Visual Tracking Via Online Discriminative Multiple Instance Metric Learning

Adaptive Resolution Optimization and Tracklet Reliability Assessment for Efficient Multi-Object Tracking

On-line Selection of Discriminative Tracking Features

Object Tracking Via Online Multiple Instance Learning with Reliable Components

Visual Object Tracking Based on Objectness Measure with Multiple Instance Learning

Robust Visual Tracking Via Multiple Discriminative Models with Object Proposals

Visual dictionary and online multi-instance learning based object tracking

Online Visual Tracking Using Multiple Instance Learning with Instance Significance Estimation.

Discriminative feature selection for visual tracking

Robust Object Tracking Via Online Multiple Instance Metric Learning.

Multiple Instance Tracking Based on Hierarchical Maximizing Bag's Margin Boosting

Online Multiple Instance Gradient Feature Selection for Robust Visual Tracking.

A Joint Object Tracking Framework with Incremental and Multiple Instance Learning

Distractor-aware discrimination learning for online multiple object tracking

Online Feature Extraction and Selection for Object Tracking

Robust Visual Tracking Via Randomly Projected Instance Learning

Multiple Instance Learning Tracking Method with Local Sparse Representation

Discriminative Object Tracking Via Sparse Representation and Online Dictionary Learning

Online Multiple Object Tracking Using a Novel Discriminative Module for Autonomous Driving