Related Experiment Videos
Automatic target recognition employing signal compression
Pradeep Ragothaman1, Wasfy B Mikhael, Robert R Muise
1School of Electrical Engineering and Computer Science, University of Central Florida, Orlando, 32816, USA.
Applied Optics
|July 5, 2007
Summary
This study introduces a novel method for automatic target recognition (ATR) using compressed Rayleigh quotient quadratic correlation filters (RQQCFs). This approach significantly reduces computational load and storage needs for infrared (IR) imagery analysis.
Area of Science:
- Computer Vision
- Signal Processing
- Infrared Imaging
Background:
- Quadratic correlation filters (QCFs) are effective for target detection and recognition in cluttered environments.
- The Rayleigh quotient quadratic correlation filter (RQQCF) was recently developed for automatic target recognition (ATR) in infrared (IR) imagery.
- RQQCF maximizes class separation metrics using training data.
Purpose of the Study:
- To present a novel approach for ATR by synthesizing RQQCFs using compressed images.
- To reduce computational complexity and storage requirements for RQQCF-based ATR.
- To maintain high recognition accuracy compared to the original RQQCF technique.
Main Methods:
- Synthesis of RQQCFs using compressed images.
- Evaluation of the proposed method on IR imagery datasets.
- Comparison of performance with the original RQQCF technique.
Main Results:
- The proposed method significantly reduces computational complexity.
- Storage requirements are considerably decreased.
- High recognition accuracy comparable to the original RQQCF is retained.
Conclusions:
- Compressed RQQCF synthesis offers an efficient alternative for ATR in IR imagery.
- The approach balances computational efficiency with high performance.
- This method is advantageous for real-time ATR applications with limited resources.