Contributions to machine learning in automotive camera and radar perception systems : applications of the sampling method in image classification and deep learning for radar-based object detection / von M. Sc. Weimeng Zhu. Wuppertal, 2023
Content
Abstract
Acknowledgement
Contents
1 Introduction
2 Fundamentals and Related Works
2.1 Fundamentals of Machine Learning
2.2 Fundamentals of Computer Vision
2.3 Fundamentals of Automotive Radar
3 A New Sampling Method for Image Classification with Neural Networks
3.1 Traffic Sign Recognition and Computer Vision Challenges
3.2 A New Sampling Method in a Convolutional Neural Network: Dense Spatial Translation Network
3.2.1 State of the Art
3.2.2 Problems and Challenges
3.2.3 DSTN Structure
3.2.4 DSTN Functionality
3.2.5 Training Strategies
3.2.6 Analysis of computational complexity
3.3 Experiment Results
3.3.1 Street View House Numbers Database
3.3.2 German Traffic Sign Recognition Benchmark Database
3.3.3 Private Traffic Sign Recognition Database
3.4 Summary
4 Radar Data Recording, Annotation, Description and Sensor Alignment
5 A Deep Neural Network for Automotive Radar Perception
5.1 State of the Art
5.2 Radar Signal Embedding with Deep Neural Network
5.3 Sensor Fusion by Deep Neural Network
5.4 A New Sampling-Based Recurrent Neural Network Module
5.4.1 Recurrent Neural Network
5.4.2 Sampling in Neural Network
5.4.3 Gated Recurrent Sampler
5.4.4 Behavior Study
5.5 A Deep Neural Network Radar Perception System
6 Conclusion
A Appendix
Bibliography
List of Figures
List of Tables
Glossary
List of Glossaries