Automatic target recognition Second Edition by Bruce Schachter ISBN 978-1510611276 1510611274
Original price was: $50.00.$35.00Current price is: $35.00.
Automatic target recognition Second Edition by Bruce J. Schachter – Ebook PDF Instant Download/Delivery: 978-1510611276, 1510611274
Full download Automatic target recognition Second Edition after payment
Product details:
ISBN 10: 1510611274
ISBN 13: 978-1510611276
Author: Bruce J. Schachter
This second edition of Automatic Target Recognition provides an inside view of the automatic target recognition (ATR) field from the perspective of an engineer working in the field for 40 years. The algorithm descriptions and testing procedures covered in the book are appropriate for addressing military problems. The book also addresses unique aspects and considerations in the design, testing, and fielding of ATR systems. These considerations need to be understood by ATR engineers working in the defense industry as well as by their government customers. The final chapter discusses the future of ATR and provides a type of Turing test for determining if an ATR system is truly smart (neuromorphic or brain-like). New to this edition is a reference design for a next-generation ATR. Coupling a Controller C to a recurrent ATR Model M forms a complete system that is more powerful in many ways than a standard ATR. This next-generation ATR can learn a never-ending sequence of tasks, adapt to unknown environments, and realize abstract planning and reasoning. It is suitable for implementation on two chips: a single, custom, low-power chip (<1 W) for implementing M, hosted by a standard processor serving as the Controller C. This ATR will be appropriate for various military systems, including those with extreme size, weight, and power constraints.
Table of contents:
1 Definitions and Performance Measures
1.1 What is Automatic Target Recognition (ATR)?
Buyers and sellers
1.2 Basic Definitions
1.3 Detection Criteria
1.4 Performance Measures for Target Detection
1.4.1 Truth-normalized measures
Assigned targets and confusers
1.4.2 Report-normalized measure
1.4.3 Receiver operating characteristic curve
1.4.4 Pd versus FAR curve
1.4.5 Pd versus list length
1.4.6 Other factors that can enter the detection equation
1.4.7 Missile terminology
1.4.8 Clutter level
1.5 Classification Criteria
1.5.1 Object taxonomy
1.5.2 Confusion matrix
1.5.2.1 Compound confusion matrix
1.5.3 Some commonly used terms from probability and statistics
1.6 Experimental Design
1.6.1 Test plan
1.6.2 ATR and human subject testing
1.7 Characterizations of ATR Hardware/Software
References
2 Target Detection Strategies
2.1 Introduction
2.1.1 What is target detection?
2.1.2 Detection schemes
2.1.3 Scale
2.1.4 Polarity, shadows, and image form
2.1.5 Methodology for algorithm evaluation
2.1.5.1 Evaluation criteria for production systems
2.1.5.2 Target detection: machine versus human
2.2 Simple Detection Algorithms
2.2.1 Triple-window filter
2.2.2 Hypothesis testing as applied to an image
2.2.3 Comparison of two empirically determined means: variations on the T-test
2.2.4 Tests involving variance, variation, and dispersion
2.2.5 Tests for significance of hot spot
2.2.6 Nonparametric tests
2.2.6.1 Percent-bright tests
2.2.7 Tests involving textures and fractals
2.2.8 Tests involving blob edge strength
2.2.9 Hybrid tests
2.2.10 Triple-window filters using several inner-window geometries
2.3 More-Complex Detectors
2.3.1 Neural network detectors
2.3.2 Discriminant functions
2.3.3 Deformable templates
2.4 Grand Paradigms
2.4.1 Geometrical and cultural intelligence
2.4.2 Neuromorphic paradigm
2.4.3 Learning-on-the-fly
2.4.4 Integrated sensing and processing
2.4.5 Bayesian surprise
2.4.6 Modeling and simulation
2.4.7 SIFT and SURF
2.4.8 Detector designed to operational scenario
2.5 Traditional SAR and Hyperspectral Target Detectors
2.5.1 Target detection in SAR imagery
2.5.2 Target detection in hyperspectral imagery
2.6 Conclusions and Future Direction
References
Appendices
3 Target Classifier Strategies
3.1 Introduction
3.1.1 Parables and paradoxes
3.2 Main Issues to Consider in Target Classification
3.2.1 Issue 1: Concept of operations
3.2.2 Issue 2: Inputs and outputs
3.2.3 Issue 3: Target classes
3.2.4 Issue 4: Target variations
3.2.5 Issue 5: Platform issues
3.2.6 Issue 6: Under what conditions does a sensor supply useful data?
3.2.7 Issue 7: Sensor issues
3.2.8
Issue 8: Processor
3.2.9 Issue 9: Conveying classification results to the human-in-the-loop
3.2.10 Issue 10: Feasibility
3.3 Feature Extraction
3.4 Feature Selection
3.5 Examples of Feature Types
3.5.1 Histogram of oriented gradients
3.5.2 Histogram of optical flow feature vector
3.6 Examples of Classifiers
3.6.1
Simple classifiers
3.6.1.1 One-class classifiers
3.6.1.2 Two-class linear classifiers
3.6.1.3 Support vector machine
3.6.2 Basic classifiers
3.6.2.1 Single-nearest-neighbor classifier
3.6.2.2 Naïve Bayes classifier
3.6.2.3 Perceptron
3.6.2.4 Learning vector quantization family of algorithms
3.6.2.5 Feedforward multilayer perceptron trained with backpropagation of error
3.6.2.6 Mean-field theory networks
3.6.2.7
Model-based classifiers
3.6.2.8 Map-seeking circuits
3.6.2.9 Ensemble classifiers
3.6.3 Contest-winning and newly popular classifiers
3.6.3.1 Hierarchical temporal memory
3.6.3.2 Long short-term memory recurrent neural network
Convolutional neural network
3.6.3.4 Sentient ATR
3.7 Discussion
References
4 Unification of Automatic Target Tracking and Automatic
Target Recognition
4.1 Introduction
4.2 Categories of Tracking Problems
4.2.1 Number of targets
4.2.2 Size of targets
4.2.3 Sensor type
4.2.4 Target type
4.3 Tracking Problems
4.3.1 Point target tracking
4.3.2 Video tracking
4.3.2.1 Correlation tracking (video data)
4.3.2.2 Feature-vector-aided tracking (video data)
4.3.2.3 Mean-shift-based moving object tracker (video tracking)
4.4 Extensions of Target Tracking
4.4.1 Activity recognition (AR)
4.4.2 Patterns-of-life and forensics
4.5 Collaborative ATT and ATR (ATT ATR)
4.5.1 ATT data useful to ATR
4.5.2 ATR data useful to ATT
4.6 Unification of ATT and ATR (ATTUATR)
4.6.1 Visual pursuit
4.6.2 A bat’s echolocation of flying insects
4.6.3 Fused ACTUATR
4.6.3.1
Spatiotemporal target detection
4.6.3.2 Forecast of features and classes
4.6.3.3 Detection-to-track association
4.6.3.4 Track maintenance
4.6.3.5 Incorporation of higher-level knowledge
4.6.3.6 Implementation
4.7 Discussion
References
5 Next-Generation ATR
5.1 Introduction
5.2 Hardware Design
5.2.1 Hardware recommendations for next-generation neuromorphic ATR
5.2.1.1 What shouldn’t be copied from biology?
5.2.1.2 What should be copied from biology?
5.3 Algorithm/Software Design
5.3.1 Classifier architecture
5.3.1.1 Decision tree
5.3.1.2 Embodied and situated (ES)
5.3.1.3 Adaptivity and plasticity (PI)
5.3.2 Embodied, situated, plastic RNN [M=ES-PI-RNN(Q)]
coupled with a controller C
5.3.2.1 Training the controller C
5.3.3 Software infrastructure
5.3.4 Test results
Contents
5.4 Potential Impact
References
6 How Smart Is Your Automatic Target Recognizer?
6.1 Introduction
6.2 Test for Determining the Intelligence of an ATR
6.2.1 Does the ATR understand human culture?
6.2.2 Can the ATR deduce the gist of a scene?
6.2.3 Does the ATR understand physics?
6.2.4 Can the ATR participate in a pre-mission briefing?
6.2.5 Does the ATR possess deep conceptual understanding?
6.2.6 Can the ATR adapt to the situation, learn on-the-fly, and make analogies?
6.2.7 Does the ATR understand the rules of engagement?
6.2.8 Does the ATR understand the order of battle and force structure?
6.2.9 Can the ATR control platform motion?
6.2.10 Can the ATR fuse information from a wide variety of sources?
6.2.11 Does the ATR possess metacognition?
6.3 Sentient versus Sapient ATR
6.4 Discussion: Where Is ATR Headed?
References
People also search for:
automatic target recognition
atr automatic target recognition
sar automatic target recognition
synthetic aperture radar automatic target recognition
ai automatic target recognition
Tags: Bruce Schachter, Automatic target