首页--工业技术论文--自动化技术、计算机技术论文--计算技术、计算机技术论文--计算机的应用论文--信息处理(信息加工)论文--模式识别与装置论文

自然场景的3D深度恢复及应用研究

摘要第4-5页
ABSTRACT第5页
Chapter 1. Introduction第10-18页
    1.1 Background第10-15页
        1.1.1 Depth Recovery From 2D Scenes第11-12页
        1.1.2 Edge Preserving Image Smoothing第12-14页
        1.1.3 RGBD Saliency Detection第14-15页
        1.1.4 Depth Data Based Online Human Action Recognition第15页
    1.2 Motivation and Contributions第15-16页
    1.3 Organization of the Dissertation第16-18页
Chapter 2. Depth Recovery From a Single Image UsingDefocus Cues第18-32页
    2.1 Introduction and Background第18-19页
    2.2 Related Work第19-21页
        2.2.1 Multiple Images Based Depth Recovery Using Defocus Cues第20页
        2.2.2 Single Image Based Depth Recovery Using Defocus Cues第20-21页
    2.3 Depth Map Recovery From a Single Image Via SpectrumContrast第21-28页
        2.3.1 Defocus Map Recovery第23-27页
        2.3.2 Depth Map Recovery第27-28页
    2.4 Experimental Results第28-31页
        2.4.1 Defocus Map Recovery Results第28-31页
        2.4.2 Depth Recovery Results第31页
    2.5 Conclusion第31-32页
Chapter 3. Edge Preserving Image Smoothing and DepthMap Refinement第32-58页
    3.1 Introduction and Background第32-33页
    3.2 Related Work第33-35页
        3.2.1 Local Filtering Based Image Smoothing Approaches第33页
        3.2.2 Optimization Based Image Smoothing Approaches第33-35页
    3.3 Our Edge Preserving Image Smoothing Algorithm第35-45页
        3.3.1 Proposed Algorithm第35-38页
        3.3.2 Textures and Edges Distinguishing第38-40页
        3.3.3 Efficient Implementation of Our Algorithm第40-43页
        3.3.4 Algorithm Analysis and Parameters Adjustment第43-45页
    3.4 Experimental Results and Comparison第45-54页
        3.4.1 Experiments on nature scenes第45页
        3.4.2 Experiments on Structured Images第45-46页
        3.4.3 Efficiency Comparison第46-48页
        3.4.4 Image Smoothing Based Applications第48-54页
    3.5 Depth Map Refinement Based on Image Smoothing第54-55页
    3.6 Depth Maps Refinement Results第55-56页
    3.7 Conclusion第56-58页
Chapter 4. Depth Assisted Saliency Detection第58-78页
    4.1 Introduction and Background第58-60页
    4.2 Related Work第60-63页
    4.3 2D Image Saliency Detection第63-68页
        4.3.1 Color Spatial Distribution (CSD) Calculation第64-65页
        4.3.2 Minimum Spanning Tree Weight (MSTW) Accumulation第65-67页
        4.3.3 Saliency Combination and Guided Filtering第67-68页
    4.4 Depth Assisted Saliency Detection第68-71页
        4.4.1 Depth Assisted Saliency Detection Model第70页
        4.4.2 Optimization and Solver第70-71页
    4.5 Experimental Results第71-75页
        4.5.1 RGB Saliency Detection Results第71-73页
        4.5.2 RGBD Saliency Detection Results第73-75页
    4.6 Conclusion第75-78页
Chapter 5. Online Human Action Recognition Based onDepth Data第78-104页
    5.1 Introduction and Background第78-80页
    5.2 Related Work第80-82页
    5.3 Skeleton Joints Based Weighted Covariance Matrices In-cremental Learning第82-92页
        5.3.1 The Covariance Descriptor第82-83页
        5.3.2 Temporal Weight and Frame Weight第83-85页
        5.3.3 Neutral Pose Model第85-88页
        5.3.4 Incremental Leaning of Weighted Covariance Matrices第88-92页
    5.4 Online Action Recognition第92-95页
        5.4.1 Online Action Recognition Using Nearest Neighbour Search第92-94页
        5.4.2 Online Action Recognition Using Log-Euclidean Kernel BasedSVM第94-95页
    5.5 Experimental Results第95-101页
        5.5.1 Skeleton Data Normalization and Denoise第96-97页
        5.5.2 Human Action Covariance Descriptor第97页
        5.5.3 Experimental Results on MSRC-12 Kinect Gesture Dataset第97-98页
        5.5.4 Experimental Results on Online Action 3D Dataset第98-100页
        5.5.5 Time Efficiency of Incrementally Updated Covariance Matrices第100-101页
    5.6 Conclusion第101-104页
Conclusion and Future Work第104-106页
Reference第106-122页
Publications and Projects第122-124页
Acknowledgement第124-125页

论文共125页,点击 下载论文
上一篇:基于局部颜色不变性特征的图像复制粘贴篡改取证研究
下一篇:基于视觉信息的质量感知模型及检索方法研究