23 Industrial Applications of Neural Networks, Neural Network, Artificial Neural Networks

[ Pobierz całość w formacie PDF ]
Industrial Applications of Neural Networks
by Lakhmi C. Jain; V. Rao Vemuri
CRC Press, CRC Press LLC
ISBN:
0849398029
Pub Date:
10/01/98
Search this book:
Chapter 1—On-Line Shape Recognition with Incremental Training Using a
Neural Network with Binary Synaptic Weights
1. Introduction
2. The Binary Synaptic Weights (BSW) Algorithm for Neural Networks
2.1 Motivation
2.2 Separation of distinct patterns by BSW algorithm
2.3 Algorithm
2.4 Example for the BSW algorithm
3. On-Line Geometric Shape Recognition with Fuzzy Filtering and Neural Network
Classification
3.1 Geometric Shape Recognition
3.2 Feature Extraction
3.2.1 Method1
3.2.1.1 Resampling
3.2.1.2 Calculation of center of the shape
3.2.1.3 Extraction of Significant Points
3.2.2 Method2
3.2.2.1 Extraction of Significant Sample Points
3.2.2.2 Calculation of the Center of Gravity
3.2.2.3 Residual Preprocessing of the Input
3.3 Formation of Tangent Vectors Along Shape Boundary
3.4 Fuzzy Function Filtering
3.5 On-line Shape Classification and Training by a BSW Network
3.6 Results
 4. Conclusion
References
Chapter 2—Neural Network Approaches to Shape from Shading
1. Introduction
2. Height and Shape Estimation
2.1 The Problem
2.2 The Forward Network Representation of a Surface
2.3 Solving the Image Irradiance Equation
2.4 Boundary Conditions
2.5 Computational Steps
2.6 Miscellaneous
3. Estimating the Illuminant Direction
4. Experiments
4.1 Examples
4.2 Several Comments
5. Extension to the Case of RBF Network
5.1 The Method
5.2 Further Examples
6. Conclusions
Acknowledgment
References
Chapter 3—Neural Networks and Fuzzy Reasoning to Detect Aircraft in
SAR Images
1. Introduction
2. Multilayer Perceptron for Classification
2.1 Structure
2.2 Classification and Training
2.3 Classification of Multispectral Images
2.4 Texture for Target Detection
3. Fuzzy Rule-Based Fusion Method
4. Experiments
5. Discussion of Results
Acknowledgments
References and Further Reading
Chapter 4—The Self-Organizing Map in Industry Analysis
1. Introduction
2. The Self-Organizing Map in Knowledge Discovery
2.1 The Self-Organizing Map
2.2 Knowledge Discovery Using the SOM
 2.2.1 Visualization
2.2.2 Clustering
2.2.3 Modeling
3. Requirements for a Forest Industry Analysis Tool
3.1 From the General to the Specific: The Neo-Generalist Between Sectoral Expert,
Amateur, and Polymath
3.2 Future Direction of Knowledge Discovery in Industry Analysis
3.3 Characteristics of the Forest Industry
3.4 Requirements for a Computerized Tool for a Forest Industry General Analyst
3.5 A SOM of ENTIRE Tasks: Mapping the Consultant’s Workload Trajectory
4. Study of Pulp and Paper Technology
4.1 Paper Machines and Pulp Lines
4.2 Mill Technology
4.3 Geographical Areas
5. Future Directions
References
Chapter 5—A Self-Organizing Architecture for Invariant 3-D Object
Learning and Recognition from Multiple 2-D Views
1. Introduction: Transforming Variant Image Data into Invariant Object Predictions
2. The Tradeoff Between Preprocessor, Classifier, and Accumulator
3. VIEWNET 1 Heuristics
4. Simulated Database
5. CORT-X 2 Filter
6. Translation, Rotation, and Scale Invariance
7. Two Coarse-Coding Strategies
8. Fuzzy ARTMAP Classification
9. Computer Simulations
9.1 Fast Learning With and Without Noise
9.2 Slow Learning Simulation With Noise
10. Voting or View Transitions for Evidence Accumulation Voting
11. Summary
Appendix A: CORT-X 2 Equations
Appendix B: Fuzzy ARTMAP Equations
References
Chapter 6—Industrial Applications of Hierarchical Neural Networks:
Character Recognition and Fingerprint Classification
1. Introduction
2. The Hierarchical Neural Network
2.1 Network Architecture
2.2 Training Algorithms
2.2.1 Self-Organizing Feature Maps
 2.2.2 Principal Component Analysis Network
2.2.3 Back-propagation Algorithm
3. Character Recognition Module of HALdoc
3.1 Multifont Character Recognition
3.2 Handwritten Digit Recognition
4. Fingerprint Classifier in HALafis
4.1 Feature Extraction
4.2 Classification of Fingerprints
4.2.1 Modified SOM Algorithm
4.2.2 Fingerprint Classification Based on UHNN
4.3 Experimental Results
5. Conclusion
Acknowledgments
References
Chapter 7—Neural Networks for Performance Optimization in Flexible
Manufacturing Systems
1. A Brief Overview of Event Graphs
1.1 Performance Optimization of Strongly Connected Event Graphs
2. FMS Performance Optimization by Event Graphs
2.1 Modeling an FMS by Event Graphs
2.2 Integer Linear Programming Problem
3. A Novel Neural Model
3.1 The Hopfield Neural Network
3.2 Description of the Novel Neural Model
3.2.1 Analytical Description of the Novel Neural Model
4. FMS Performance Optimization Using the Novel Neural Model
4.1 Modeling the FMS by an Event Graph
4.2 Mapping the Event Graph onto the Neural Model
4.3 Determination of Weights and Bias Currents for the Novel Neural Model
4.3.1 Quality of the Solution
4.3.2 Validity of the Solution
4.4 FMS Performance Optimization by the Novel Neural Model
5. Examples of FMS Performance Optimization by the Neural Approach
6. Some Considerations on the Neural Computation
7. Conclusion
References
Chapter 8—Channel Assignment in Mobile Communication Networks - A
Computational Intelligence Approach
1. Introduction
2. The Channel Assignment Problem
3. Problem Formulation
 4. Convergence and Determination of the Frequency Span
4.1 Case 1
4.2 Case 2
4.3 Convergence of the MFA Channel Assignment Algorithm
5. Numerical Examples and Conclusions
Appendix A: Mean Field Annealing
A.1 Statistical Mechanics
A.2 Simulated Annealing
A.3 Mean Field Annealing
A.4 Convergence of MFA
Appendix B: NP-Completeness of the Channel Assignment Problem
References
Chapter 9—Application of Cellular Compact Neural Networks in Digital
Communication
1. Introduction
2. Cellular/Compact Neural Networks
2.1 Basic Theory and Computation Paradigm
2.2 Stability
3. 1-D Compact Neural Networks for Wireless Communication
3.1 Digital Communication and Neural Networks
3.2 System Model
3.2.1 Combinatorial Optimization
3.2.2 Cloning Templates
3.3 Hardware Annealing
3.4 Simulation Results
4. 1-D Compact Neural Network for Communication: Applications
4.1 Partial Response Maximum Likelihood (PRML) Sequence Detector
4.1.1 State-Constrained Neuron Model
4.1.2 Maximum Likelihood Sequence Detection (MLSD) Algorithm
4.1.3 Performance of Algorithm
4.2 A CDMA Communication Detector with Robust Near-Far Resistance
4.2.1 Conventional Detector and Optimized Decision Rule
4.2.2 Compact Neural Network CDMA Detector
4.2.3 Simulation Results
4.3 Conclusion
Acknowledgments
References
Chapter 10—Neural Networks for Process Scheduling in Communication
Systems
1. Characteristics of the Processes in a Communication System
2. Process Scheduling in a Communication System
  [ Pobierz całość w formacie PDF ]
  • zanotowane.pl
  • doc.pisz.pl
  • pdf.pisz.pl
  • charloteee.keep.pl