Abstract:Traditional tracking methods place an emphasis on how to cope with the variations in target appearance effectively. However, when the motion displacement of the target between image frames becomes larger, these methods may be unstable. This paper presents a novel (to our knowledge) visual object tracking method. In this method, we first introduce scale-invariant feature transform (SIFT) flow into the tracking problem and develop a real-time motion prediction method to capture large displacement between consecutive image frames. Then we use belief propagation (BP) to convert the problem of finding maximum a posteriori probability (MAP) to globally minimizing an energy function to get the best matching pairs of points for producing good candidate regions of the target. And last, the refined point trajectories are obtained according to the bidirectional flow field consistency estimation and covariance region descriptor matching, which can update model states efficiently so as to achieve enhanced robustness for visual tracking. Compared with the state-of-art tracking methods, the experimental results demonstrate that the proposed algorithm shows favorable performance when the object undergoes large motion displacement between image frames.

Sift Features Based Object Tracking With Discrete Wavelet Transform

The Research of Video Tracking Based on Improved Sift Algorithm

Tracking Object by Combining Particle Filters and SIFT Features

Moving Object Detection Based on SIFT Features Matching and K-means Clustering

Robust Object Tracking Combining Color and Scale Invariant Features

Optical flow motion tracking algorithm based on SIFT feature

Multi-objects Tracking and Online Identification Based on SIFT

SIFT feature tracking in multi-beam sonar image sequences using the track-before-detect technique

Efficient modified directional lifting-based discrete wavelet transform for moving object detection

SIFT Flow for Large-Displacement Object Tracking

A Method of Specific Moving Objects Detection and Tracking in Complex Background

Traffic Object Detection and Tracking Based on Wavelet Lifting Scheme

A Selective Tracking and Detection Framework with Target Enhanced Feature

Object Tracking Algorithm Based on Combination of Dynamic Template Matching and Kalman Filter

A moving object tracking method based on sequential detection scheme

Object Tracking with Kalman Filter and Discrete Cosine Transform

Target Tracking Method Based on SURF

The Accurate Extraction and Tracking of Moving Objects for Video Surveillance

A Novel Object Tracking Algorithm Based On Enhanced Perception Hash And Online Template Matching

A New Object Tracking Algorithm Based on the Fast Discrete Curvelet Transform

An Approach for Object Matching Based on SIFT Feature in Multiple-Camera Surveillance System