首页--工业技术论文--自动化技术、计算机技术论文--自动化基础理论论文--人工智能理论论文--自动推理、机器学习论文

深度学习算法在无人驾驶视觉中的应用

Abstract第5-7页
摘要第8-14页
List of Abbreviations第14-23页
Chapter 1 Introduction第23-35页
    1.1 Background第23-29页
        1.1.1 Introduction of Self-driving Vehicle第23-25页
        1.1.2 Visual Perception In Self-driving Vehicle System第25-26页
        1.1.3 Road Detection第26-28页
        1.1.4 Object Detection and Recognition第28-29页
    1.2 Problems and Challenges第29-32页
        1.2.1 Road Detection第29-31页
        1.2.2 Object Detection and Recognition第31-32页
    1.3 Dissertation Outline第32-35页
Chapter 2 Fundamentals of Deep Learning第35-45页
    2.1 Introduction第35-36页
    2.2 Neural Network in Early Stage第36-40页
        2.2.1 Multilayer Perception第36-37页
        2.2.2 Restricted Boltzmann Machine第37-40页
    2.3 Convolutional Neural Network第40-44页
        2.3.1 Basic Modules in CNN structures第40-42页
        2.3.2 Typical CNN-Based Models for Object Recognition第42页
        2.3.3 Platforms for CNN Model Development第42-44页
    2.4 Conclusion第44-45页
Chapter 3 CNN-based Road-direction Point Detection第45-69页
    3.1 Introduction第45-47页
    3.2 Proposed Method第47-55页
        3.2.1 Road-direction Point Representation第47-49页
        3.2.2 Design of Road-direction Point Detection Model第49-50页
        3.2.3 Loss function第50-51页
        3.2.4 Convolutional Neural Network Structure第51-53页
        3.2.5 Non-maximum suppression第53页
        3.2.6 Training of This Model第53-55页
    3.3 Simulation Results第55-67页
        3.3.1 Design of Dataset About Road第55-58页
        3.3.2 Model Simulation第58-59页
        3.3.3 Performance Evaluation第59-64页
        3.3.4 Performance Comparison第64-66页
        3.3.5 Runtime Comparison第66-67页
    3.4 Discussion第67-68页
    3.5 Conclusion第68-69页
Chapter 4 CNN-based Multiple Road-Points Detection第69-93页
    4.1 Introduction第69-71页
    4.2 Proposed Method第71-80页
        4.2.1 Road Representation by Road Points第71-73页
        4.2.2 Design of Road-Points Detection Model第73-74页
        4.2.3 Loss Function第74-76页
        4.2.4 Convolutional Neural Network Structure第76-77页
        4.2.5 Non-maximum Suppression第77-78页
        4.2.6 Training of This Model第78页
        4.2.7 Metric Definition for Model Performance第78-80页
    4.3 Simulation Results第80-91页
        4.3.1 Design of Road Dataset第80页
        4.3.2 Model Simulation第80-82页
        4.3.3 Error Analysis第82-84页
        4.3.4 Mean Performance on Different Categories第84-85页
        4.3.5 Performance Comparison第85-90页
        4.3.6 Runtime Comparison第90-91页
    4.4 Conclusion第91-93页
Chapter 5 Road-direction Point based Car Detection第93-117页
    5.1 Introduction第93-95页
    5.2 Proposed Method第95-105页
        5.2.1 CNN-based Model with Sub-regions第95-99页
        5.2.2 Information Integration第99-101页
        5.2.3 Loss Function第101-102页
        5.2.4 Convolutional Neural Network Structure第102-104页
        5.2.5 Training of this model第104-105页
    5.3 Simulatoin Results第105-114页
        5.3.1 Preparation of Dataset第105-106页
        5.3.2 Model Simulation第106页
        5.3.3 Model Performance第106-112页
        5.3.4 Analysis of Model Performance第112-114页
        5.3.5 Runtime Comparison第114页
    5.4 Conclusion第114-117页
Chapter 6 Research on Invariant Object Recognition第117-137页
    6.1 Introduction第117-118页
    6.2 Proposed Method第118-125页
        6.2.1 Sparse Deep Belief Network第118-120页
        6.2.2 V2 Features Detection with 2-Stage DBN第120-121页
        6.2.3 SOM Model with Trace Rule第121-124页
        6.2.4 Metric Method for Model Performance第124-125页
    6.3 Simulation Results第125-134页
        6.3.1 Design of Simulation第125-127页
        6.3.2 Influence of Trace Rule on Performance第127-129页
        6.3.3 Influence of the Number of SOM Layers第129-130页
        6.3.4 Influence of Random Order on Performance第130-131页
        6.3.5 Comparison of Firing Rate of SSI-Top neurons第131-133页
        6.3.6 Application of Learned SOM Layer第133-134页
    6.4 Discussion第134-135页
    6.5 Conclusion第135-137页
Chapter 7 Conclusions and Future Works第137-141页
    7.1 Conclusions第137-140页
    7.2 Future Works第140-141页
Appendix第141-143页
References第143-153页
Acknowledgements第153-155页
Biography第155页

论文共155页,点击 下载论文
上一篇:当代大学生中国特色社会主义道路认同教育研究
下一篇:普惠金融与减贫:不同模式及对马里选择的意义