Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Video

Updated: Jun 12, 2026

Utilizing vmTracking to Improve the Accuracy of Multi-Animal Pose Estimation in Rodent Social Behavior Studies

Utilizing vmTracking to Improve the Accuracy of Multi-Animal Pose Estimation in Rodent Social Behavior Studies

Published on: November 7, 2025

Online empirical evaluation of tracking algorithms.

Hao Wu¹, Aswin C Sankaranarayanan, Rama Chellappa

¹Center for Automation Research, University of Maryland, College Park, MD 20742, USA. wh2003@umiacs.umd.edu

IEEE Transactions on Pattern Analysis and Machine Intelligence

|June 19, 2010

Summary

This study introduces a novel online method to evaluate visual tracking systems without ground truth. The approach uses a time-reversed Markov chain to detect tracking failures caused by occlusion or model errors.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Computational Imaging for Long-Term Prediction of Solar Irradiance.

IEEE transactions on pattern analysis and machine intelligence·2025

Same author

Spatial and axial resolution limits for mask-based lensless cameras.

Optics express·2023

Same author

Wide-Baseline Light Fields Using Ellipsoidal Mirrors.

IEEE transactions on pattern analysis and machine intelligence·2022

Same author

PS$^{2}$2 F: Polarized Spiral Point Spread Function for Single-Shot 3D Sensing.

IEEE transactions on pattern analysis and machine intelligence·2022

Same author

Designing Display Pixel Layouts for Under-Panel Cameras.

IEEE transactions on pattern analysis and machine intelligence·2021

Same author

SASSI - Super-Pixelated Adaptive Spatio-Spectral Imaging.

IEEE transactions on pattern analysis and machine intelligence·2021

Area of Science:

Computer Vision
Machine Learning
Signal Processing

Background:

Evaluating visual tracking systems without ground truth is a significant challenge.
Existing methods often struggle to scale with complex, high-dimensional visual tracking models.
Diverse scenarios like occlusion and illumination changes complicate performance assessment.

Purpose of the Study:

To propose an online performance evaluation strategy for tracking systems, specifically those using particle filters.
To develop a method for detecting tracking failures by leveraging the time-reversible nature of physical motion.
To provide a robust evaluation metric applicable even with occlusion, pose, and illumination variations.

Main Methods:

Utilizing a time-reversed Markov chain initialized with the tracker's posterior distribution.

More Related Videos

Eye-tracking Technology and Data-mining Techniques used for a Behavioral Analysis of Adults engaged in Learning Processes

Eye-tracking Technology and Data-mining Techniques used for a Behavioral Analysis of Adults engaged in Learning Processes

Published on: June 10, 2021

Related Experiment Videos

Last Updated: Jun 12, 2026

Utilizing vmTracking to Improve the Accuracy of Multi-Animal Pose Estimation in Rodent Social Behavior Studies

Utilizing vmTracking to Improve the Accuracy of Multi-Animal Pose Estimation in Rodent Social Behavior Studies

Published on: November 7, 2025

Eye-tracking Technology and Data-mining Techniques used for a Behavioral Analysis of Adults engaged in Learning Processes

Eye-tracking Technology and Data-mining Techniques used for a Behavioral Analysis of Adults engaged in Learning Processes

Published on: June 10, 2021

Computing the posterior density by filtering backward in time to the initial instant.

Employing the distance between the time-reversed posterior and the prior as a decision statistic for failure detection.

Main Results:

Demonstrated effectiveness in detecting tracking failures due to occlusion, pose, and illumination changes.
Achieved low decision statistic values when tracking data aligns with underlying models.
Presented Receiver Operating Characteristic (ROC) curves to validate performance.

Conclusions:

The proposed time-reversed Markov chain method offers a scalable and effective solution for online evaluation of visual tracking systems without ground truth.
The methodology successfully identifies tracking failures by detecting violations in the expected time-reversible nature of motion.
The approach is adaptable to various tracking algorithms, including Kanade-Lucas-Tomasi (KLT) and mean-shift trackers.