Video Content Analysis Using Multimodal Information

Dedication. List of Figures. List of Tables. Preface. Acknowledgments. 1: Introduction. 1. Audiovisual Content Analysis. 1.1. Audio Content Analysis. 1.2. Visual Content Analysis. 1.3. Audiovisual Content Analysis. 2. Video Indexing, Browsing and Abstraction. 3. MPEG-7 Standard. 4. Roadmap of The Book. 4.1. Video Segmentation. 4.2. Movie Content Analysis. 4.3. Movie Content Abstraction. 2: Background And Previous Work. 1. Visual Content Analysis. 1.1. Video Shot Detection. 1.2. Video Scene and Event Detection. 2. Audio Content Analysis. 2.1. Audio Segmentation and Classification. 2.2. Audio Analysis for Video Indexing. 3. Speaker Identification. 4. Video Abstraction. 4.1. Video Skimming. 4.2. Video Summarization. 5. Video Indexing and Retrieval. 3: Video Content Pre-Processing. 1. Shot Detection in Raw Data Domain. 1.1. YUV Color Space. 1.2. Metrics for Frame Differencing. 1.3. Camera Break Detection. 1.4. Gradual Transition Detection. 1.5. Camera Motion Detection. 1.6. Illumination Change Detection. 1.7. A Review of the Proposed System. 2. Shot Detection in Compressed Domain. 2.1. DC-image and DC-sequence. 3. Audio Feature Analysis. 4. Commercial Break Detection. 4.1. Features of A Commercial Break. 4.2. Feature Extraction. 4.3. The Proposed Detection Scheme. 5. Experimental Results. 5.1. Shot Detection Results. 5.2. Commercial Break Detection Results. 4: Content-Based Movie Scene And Event Extraction. 1. Movie Scene Extraction. 1.1. Sink-based Scene Construction. 1.2. Audiovisual-based Scene Refinement. 1.3. User Interaction. 2. Movie Event Extraction. 2.1. Sink Clustering and Categorization. 2.2. Event Extraction and Classification. 2.3. Integrating Speech and Face Information. 3. Experimental Results. 3.1. Scene Extraction Results. 3.2. Event Extraction Results. 5: Speaker Identification For Movies. 1. Supervised Speaker Identification for Movie Dialogs. 1.1. Feature Selection and Extraction. 1.2. Gaussian Mixture Model. 1.3. Likelihood Calculation and Score Normalization. 1.4. Speech Segment Isolation. 2. Adaptive Speaker Identification. 2.1. Face Detection, Recognition and Mouth Tracking. 2.2. Speech Segmentation and Clustering. 2.3. Initial Speaker Modeling. 2.4. Likelihood-based Speaker Identification. 2.5. Audiovisual Integration for Speaker Identification. 2.6. Unsupervised Speaker Model Adaptation. 3. Experimental Results. 3.1. Supervised Speaker Identification Results. 3.2. Adaptive Speaker Identification Results. 3.3. An Example of Movie Content Annotation. 6: Scene-Based Movie Summarization. 1. An Overview of the Proposed System. 2. Hierarchical Keyframe Extraction. 2.1. Scene Importance Computation. 2.2. Sink Importance Computation. 2.3. Shot Importance Computation. 2.4. Frame Importance Computation. 2.5. Keyframe Selection. 3. Scalable Movie Summarization and Navigation. 4. Experimental Results. 4.1. Keyframe Extraction Results. 4.2. User Study. 4.3. System Interface Design. 4.4. Applications. 7: Event-Based Movie Skimming. 1. Introduction. 2. An Overview of the Proposed System. 3. Extended Event Set Construction. 4. Extended Event Feature Extraction. 5. Video Skim Generation. 6. More Thoughts on the Video Skim. 6.1. When More Judging Rules Are Needed. 6.2. Sub-sampling the Video Skim. 6.3. Discovering the Story and Visual Structure. 7. Experimental Results. 8: Conclusion And Future Work. 1. Conclusion. 2. Future Work. 2.1. System Refinement. 2.2. New Research Topics. References. Index.