第一个书签之前
基于明星的视频人脸验证和聚类算法研究
RESEARCH ON SITCOM-STAR ORIENTED FACE VERIFICATION AND CLUSTERING FOR VIDEOS
摘 要
Abstract
Acknowledgement
Contents
Nomenclature
List of Tables
List of Figures
Chapter 1 Introduction
1.1 Background, purpose and significance of this research
1.2 Examination of research status and standard procedures
1.2.1 Research on face detection
1.2.2 Research on face clustering
1.3 The main content of this research
Chapter 2 Literature Review
2.1 Face detection
2.1.1 Facial landmark localization
2.1.2 Face normalization
2.2 Face tracking
2.2.1 Metric learning for face trackers
2.2.2 Metric for identification
2.3 Face feature extraction
2.3.1 Local binary patterns
2.4 Summary
Chapter 3 FACE VERIFICATION AND CLUSTERING ON SITCOM-STAR ORIENTED VIDEOS
3.1 System design
3.2 Feature extraction and face verification
3.3 Face clustering on videos
3.3.1 Spectral clustering based method
3.3.2 Sparse subspace clustering method
3.3.3 Density peaks clustering
3.3.4 An improved distance measurement for clustering algorithms
3.3.5 Face clustering by using key frame shots
3.4 Summary
Chapter 4 Face Clustering on videos and experiments
4.1 Experimental setup
4.2 Dataset overview
4.3 System modules
4.3.1 Face detection and face alignment
4.3.2 Face verification
4.3.3 Problem formulation of sitcom-star face clustering
4.3.4 Clustering by finding density peaks
4.3.5 The density peaks clustering algorithm
4.4 Experimental results for sitcom-star face clustering
4.4.1 Face image dataset
4.4.2 Evaluation metric
4.4.3 Data visualization and experimental results
4.5 Summary
Conclusions
References
哈尔滨工业大学学位论文原创性声明和使用权限