This course focuses on introducing a new framework for multi-target tracking using a data association technique based on the Generalized Maximum Clique Problem (GMCP) formulation. The course covers the limitations of GMCP and introduces the Generalized Maximum Multi Clique Problem (GMMCP) to address these limitations. It also discusses the challenges of tracking a large number of targets in crowded scenarios and proposes a solution using Binary Quadratic Programming. The course teaches skills in data association, graph theoretic problem formulation, optimization, and the application of the Frank-Wolfe algorithm. The teaching method includes theoretical explanations, problem-solving demonstrations, and comparisons of different tracking methods. This course is intended for individuals interested in computer science, specifically in the field of multi-target tracking and data association.
Overview
Syllabus
Intro
Challenges
Multi-target Tracking: Applications
Outline
Data Association
GMCP Tracker: Pipeline
How to solve GMCP?
Process of Finding Tracklets in one Segment
Parking Lot Results
Evaluation Metrics
Limitations
What are the main differences?
Framework
Mid-level Tracklet Generation
Optimization
Aggregated Dummy Nodes (ADN)
Run-time Comparison
Qualitative Results
Parking Lot 2
Occlusion Handling
Quantitative Comparison
Crowd Tracking
Spatial Proximity Constraint
Neighborhood Motion Effect
Grouping
Formulation
Appearance
Quadratic Constraints
Frank Wolfe Algorithm
Frank Wolfe with SWAP steps
Experiments . 9 high-density sequences
Quantitative Results
Contribution of each term
Summary
Future Work
Taught by
UCF CRCV
