Content parsing of home videos by motion analysis

Due to the increasing use of hand-held camcorders, an explosion of home video data is already underway. Home videos, by nature, are unedited, unstructured and lack of story-line. They contain unrestricted content domain which usually mixes together with irregular camera motions. These features have made conventional video processing techniques inappropriate for the analysis of home videos. In this thesis, we propose new techniques for the automatic parsing of home video content to support effective indexing and browsing. These techniques cover two major issues: video object analysis (VOA) and content parsing (CP). In VOA, a new algorithm is proposed for automatic object initialization. This algorithm is based on motion discriminant analysis formulated through 3D tensor representation and robust clustering…


1 Introduction
1.1 Motivation and Objectives
1.1.1 The Proposed Framework
1.2 Contributions
1.3 Thesis Organization
2 Background
2.1 Background Introduction
2.2 Related Work
2.2.1 Home Video Processing
2.2.2 Object Detection and Tracking
2.2.3 Object Initialization
2.2.4 Video Stabilization
3 Multiple Object Initialization
3.1 Introduction
3.2 Seed Parametrization
3.3 Seed Candidate Detection
v3.3.1 3D tensor representation
3.3.2 Motion segmentation and seed candidate detection
3.4 Seed Association
3.4.1 Seed Dynamical Model
3.4.2 Seed Candidate Model
3.4.3 Seed Kalman Filtering
3.5 Temporal Seed Selection
3.5.1 Approach I: Single Selection
3.5.2 Approach II: Sequential Selection
3.5.3 Approach III: Synchronous Selection
3.6 Experiments
3.7 Summary
4 Object Detection and Tracking
4.1 Introduction
4.2 EM-based Object Detection
4.3 Meanshift Object Tracking
4.4 Experiments
4.5 Summary
5 Snippet Detection and Pattern Parsing
5.1 Introduction
5.2 Constructing Table-of-Snippet
5.2.1 3D Tensor Representation
5.2.2 Motion Feature Extraction
5.2.3 Zoom Segment Detection
5.2.4 Polyline Estimation
5.2.5 Snippet Boundary Detection
5.2.6 Keyframe Representation
5.3 Pattern Matching by MWBG
5.4 Experiments
5.4.1 Snippet Boundary Detection
vi5.4.2 Pattern Parsing and Indexing
5.5 Summary
6 Motion Stabilization
6.1 Introduction
6.1.1 Overview of Our Approach
6.2 Motion Characterization
6.2.1 3D Tensor Representation
6.2.2 Motion Clustering
6.3 Motion Segmentation
6.4 Video Stabilization
6.4.1 Kalman Filter
6.4.2 Selective Stabilizers
6.5 Experiments
6.6 Summary
7 Conclusion and Future Work
7.1 Conclusion
7.2 Future Work
7.2.1 Multiple Object Initialization
7.2.2 Object Analysis
7.2.3 Event Analysis

Author: Pan, Zailiang

Source: City University of Hong Kong

Download URL 2: Visit Now

Leave a Comment