An Experiment on Target Tracking via Image Segmentation.
Abstract
A target tracking and detection experiment is reported that uses the Fisher's linear discriminant for pixel classification to segment simulated static and dynamic scenes. The results clearly demonstrate that the segmentation method performs better than other work on the same static scene. Detection performance versus the learning sample sizes for dynamic scenes is empirically determined. It indicates that a small target can still be detected if a sufficiently large learning sample size is available. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 11, 1982
- Accession Number
- ADA109618
Entities
People
- Chia‐Hung Chen
- Wen-hsing Yang
Organizations
- University of Massachusetts Dartmouth