COURSE # ROO-421
OPTICAL SIGNAL PROCESSING AND PATTERN RECOGNITION
Comprehensive introduction and an overview of optical signal processing and its applications
in pattern recognition.
Optical systems find many well known and attractive applications in sensing, recording, storage and transmission of data. These applications stimulated the development of many new optical devices, components, materials and applied technologies. This course considers the use of optical systems to process data, rather then merely for sensing, storage or fiber communications. The specific application considered is pattern recognition with special attention paid to the practical and difficult problems of distortion-invariant pattern recognition of low contrast objects in the presence of clutter. These conditions are of special interest in automatic target recognition, robotics, machine and computer vision, character recognition, etc.
Applications and benefits:
You will benefit by enhancing your understanding of the :
- Fundamental concepts of optics, linear systems, Fourier transforms and classifiers.
- Linear algebra and optimization concepts used in pattern recognition.
- New algorithms for optical and digital processing.
- Basic operations, optical architectures and algorithms required in pattern recognition.
- Optical building blocks and the design and fabrication of optical processing systems.
- 2-D and 1-D signal processing optical architectures.
Who should attend:
This course introduces the principles and techniques used in optical signal processing and is presented in the context of its applications to optical pattern recognition. The basics, current state-of-the-art, and future trends and needs presented render this course an invaluable resource for managers, system analysts, engineers and designers involved in pattern recognition projects and optical signal processing. This course has no prerequisites; however, a background in engineering, science or equivalent experience will be helpful.
Course Outline:
- Overview
- Operations desired
- Introduction to linear systems
- Fourier transforms
- Convolution and correlation
- Shift-invariance
- Introduction to optics
- Ray optics, wave optics
- Coherence, diffraction, interference
- Impulse response, transfer function
- Optical Fourier Transform
- Math, architectures
- Space bandwidth, frequency resolution
- Design and fabrication
- Optical filtering
- Amplitude, phase, directional
- Edge enhancement, image enhancement
- Optical correlators
- Matched spatial filters
- Pattern recognition, holography
- Architectures, design and fabrication
- Components
- Computer generated holograms
- Spatial light modulators
- Hardware, architectures
- Data
- Visible, infrared
- Synthetic aperture radar
- 3D range
- Classifiers
- Linear algebra
- Discrimination functions
- Feature extractors
- Examples, case studies
- Space variant
- Object identification
- Product inspection
- Operations achievable (and architectures)
- Fourier transform, correlation
- Filtering, feature generation
- Morphological processors
- Wavelet and Gabor processor
- Distortion-invariant filters
- Neural networks
- Fusion
- Detection, recognition, and classification
- 1-D signal processing
- Range and Doppler concepts
- Acousto-Optic processor
Text: Real Time Optical Information Processing, B.B. Javidi and J. Horner, eds.; Academic Press, 1994.
About the Instructor
Professor Casasent is a Full Professor at Carnegie Mellon University, Pittsburgh, PA, in the Department of Electrical and Computer Engineering, where he is the George Westinghouse Professor and Director of the Laboratory for Optical Data Processing. He is a Fellow of the IEEE, OSA and SPIE and has received various best paper awards and other honors. He is the author of two books, editor of one text, editor of 50 journals and conference volumes, contributor to chapters in 20 books and over 600 technical publications, on various aspects of optical data processing, image pattern recognition, and real-time signal processing. Dr. Casasent is active in conference organizations and is a consultant to companies and government agencies. He originated and has organized the set of 6 to 11 annual SPIE conferences on Intelligent Robots and Computer Vision. He is Past President of SPIE and was on the Board of Directors of SPIE for 6 years. He is a past member of two Defense Science Board Task Forces (on Image Recognition and on Automatic Target Recognition). He is Past President of the Pittsburgh chapters of the IEEE-ED and the Optical Society of America. He is presently Faculty Advisor to Eta Kappa Nu among other such activities. His research interests include: distortion-invariant pattern recognition, neural networks, Gabor and wavelet transforms, and morphological image processing.
Professor Casasent is presently on the Board of Directors of INNS (the International Neural Network Society) and is an Editorial Board member of the Neural Networks journal and six other journals.
Details:
Course: ROO-421 Duration: 3 Days FEE: $1,195 CEUs: 2.16
Please direct any additional inquiries regarding this course to Robert Blakely, Program Director, by e-mail, FAX: (301) 871-4942 or TELEPHONE: (301) 871-9608.
Call toll free 1-800-683-7267 from anywhere in the Continental U.S. or CANADA.
Last modified December 12, 1997.
