Authors:
Sławomir Bąk
1
;
Sundaram Suresh
2
;
François Brémond
2
and
Monique Thonnat
2
Affiliations:
1
Institute of Computing Science, Poznan University of Technology, Poland
;
2
INRIA Sophia Antipolis, PULSAR group, France
Keyword(s):
Object tracking, Neural network, Gaussian activation function, Feature extraction, On-line learning, Motion segmentation, Reliability classification.
Related
Ontology
Subjects/Areas/Topics:
Applications
;
Artificial Intelligence
;
Biomedical Engineering
;
Biomedical Signal Processing
;
Computational Intelligence
;
Computer Vision, Visualization and Computer Graphics
;
Detecting 3D Objects Using Patterns of Motion and Appearance
;
Feature Extraction
;
Features Extraction
;
Health Engineering and Technology Applications
;
Human-Computer Interaction
;
Image and Video Analysis
;
Informatics in Control, Automation and Robotics
;
Methodologies and Methods
;
Model-Based Object Tracking in Image Sequences
;
Motion and Tracking
;
Motion, Tracking and Stereo Vision
;
Neural Networks
;
Neurocomputing
;
Neurotechnology, Electronics and Informatics
;
Pattern Recognition
;
Physiological Computing Systems
;
Retrieval of 3D Objects from Video Sequences
;
Sensor Networks
;
Signal Processing
;
Signal Processing, Sensors, Systems Modeling and Control
;
Soft Computing
;
Theory and Methods
;
Tracking of People and Surveillance
Abstract:
This paper presents a method to fuse the information from motion segmentation with online adaptive neural classifier for robust object tracking. The motion segmentation with object classification identify new objects present in the video sequence. This information is used to initialize the online adaptive neural classifier which is learned to differentiate the object from its local background. The neural classifier can adapt to illumination variations and changes in appearance. Initialized objects are tracked in following frames using the fusion of their neural classifiers with the feedback from the motion segmentation. Fusion is used to avoid drifting problems due to similar appearance in the local background region. We demonstrate the approach in several experiments using benchmark video sequences with different level of complexity.