Embedded Computer Vision

Foreword

Preface

Embedded Computer Vision

Target Audience

Organization of the Book

Overview of Chapters

How This Book Came About

Outlook

Acknowledgements

Contents

List of Contributors

Introduction

Hardware Considerations for Embedded Vision Systems

1.1 The Real-Time Computer Vision Pipeline

1.2 Sensors

1.3 Interconnects to Sensors

1.4 Image Operations

1.5 Hardware Components

1.6 Processing Board Organization

1.7 Conclusions

References

Design Methodology for Embedded Computer Vision Systems

2.1 Introduction

2.2 Algorithms

2.3 Architectures

2.4 Interfaces

2.5 Design Methodology

2.6 Conclusions

References

We Can Watch It for You Wholesale

3.1 Introduction to Embedded Video Analytics

3.2 Video Analytics Goes Down-Market

3.3 How Does Video AnalyticsWork?

3.4 An Embedded Video Analytics System: by the Numbers

3.5 Future Directions for Embedded Video Analytics

3.6 Conclusion

References

Advances in Embedded Computer Vision

Using Robust Local Features on DSP-Based Embedded Systems

4.1 Introduction

4.2 RelatedWork

4.3 Algorithm Selection

4.4 Experiments

4.5 Conclusion

References

Benchmarks of Low-Level Vision Algorithms for DSP, FPGA, and Mobile PC Processors

5.1 Introduction

5.2 RelatedWork

5.3 Benchmark Metrics

5.4 Implementation

5.5 Results

5.6 Conclusions

References

SAD-Based Stereo Matching Using FPGAs

6.1 Introduction

6.2 RelatedWork

6.3 Stereo Vision Algorithm

6.4 Hardware Implementation

6.5 Experimental Evaluation

6.6 Conclusions

References

Motion History Histograms for Human Action Recognition

7.1 Introduction

7.2 RelatedWork

7.3 SVM-Based Human Action Recognition System

7.4 Motion Features

7.5 Dimension Reduction and Feature Combination

7.6 System Evaluation

7.7 FPGA Implementation on Videoware

7.8 Conclusions

References

Embedded Real-Time Surveillance Using Multimodal Mean Background Modeling

8.1 Introduction

8.2 RelatedWork

8.3 Multimodal Mean Background Technique

8.4 Experiment

8.5 Results and Evaluation

8.6 Conclusion

References

Implementation Considerations for Automotive Vision Systems on a Fixed- Point DSP

9.1 Introduction

9.2 Fixed-Point Arithmetic

9.3 Process of Dynamic Range Estimation

9.4 Implementation Considerations for Single-Camera Steering Assistance Systems on a Fixed- Point DSP

9.5 Results

9.6 Conclusions

References

Towards OpenVL: Improving Real-Time Performance of Computer Vision Applications

10.1 Introduction

10.2 RelatedWork

10.3 A Novel Software Architecture for OpenVL

10.4 Example Application Designs

10.5 Conclusion and Future Work

10.6 Acknowledgements

References

Looking Ahead

Mobile Challenges for Embedded Computer Vision

11.1 Introduction

11.2 In Search of the Killer Applications

11.3 Technology Constraints

11.4 Intangible Obstacles

11.5 Future Direction

References

Challenges in Video Analytics

12.1 Introduction

12.2 Current Technology and Applications

12.3 Building Blocks

12.4 Embedded Implementations

12.5 Future Applications and Challenges

12.6 Summary

References

Challenges of Embedded Computer Vision in Automotive Safety Systems

13.1 Computer Vision in Automotive Safety Applications

13.2 Literature Review

13.3 Vehicle Cueing

13.4 Feature Extraction

13.5 Feature Selection and Classification

13.6 Experiments

13.7 Conclusion

References

Index