Visual tracking using superpixel-based appearance model

Authors:
Shahed Nejhum;Muhammad Rushdi;Jeffrey Ho
Affiliations:
The MathWorks, Natick, MA;Department of Computer and Information Science and Engineering, University of Florida, Gainesville, FL;Department of Computer and Information Science and Engineering, University of Florida, Gainesville, FL
Venue:
ICVS'13 Proceedings of the 9th international conference on Computer Vision Systems
Year:
2013

Citing 8
Cited 0

Color indexing

International Journal of Computer Vision
Robust Fragments-based Tracking using the Integral Histogram

CVPR '06 Proceedings of the 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition - Volume 1
Incremental Learning for Robust Visual Tracking

International Journal of Computer Vision
Semi-supervised On-Line Boosting for Robust Tracking

ECCV '08 Proceedings of the 10th European Conference on Computer Vision: Part I
TurboPixels: Fast Superpixels Using Geometric Flows

IEEE Transactions on Pattern Analysis and Machine Intelligence
Superpixels and supervoxels in an energy optimization framework

ECCV'10 Proceedings of the 11th European conference on Computer vision: Part V
Robust Object Tracking with Online Multiple Instance Learning

IEEE Transactions on Pattern Analysis and Machine Intelligence
Superpixel tracking

ICCV '11 Proceedings of the 2011 International Conference on Computer Vision

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this work, we propose a tracking algorithm that robustly handles complex variations in target appearance, scale, occlusion, and background. In particular, the algorithm exploits a novel superpixel-based appearance model for visual tracking. From the initial tracking window, we extract superpixels and compute their histogram features. In subsequent frames, we search for the region that maximizes the similarity of the superpixel features. Our algorithm detects target occlusion and updates the appearance model accordingly. As well, the model is updated to handle large-scale variations. We present experimental results on several publicly available challenging sequences. Qualitative and quantitative evaluation of our tracking algorithm show improved performance over state-of-the-art trackers.