Person Re-Identification and Tracking in Multiple Non-Overlapping Cameras - Keynote
University of Central Florida via YouTube
Overview
Syllabus
Intro
Challenges
Person Classification Network
SPREID - Human Semantic Parsing
Triplet Loss Based Network
Quadruplet Loss Based Network
Diffusion Based Network
DCDS Based Network
Pipeline
Dominant Sets Clustering
Constrained Dominant Sets (CDS)
Auxiliary Net
At Testing
Constraint Expansion
Different Camera Configurations For Tracking
Multiple Fixed & Overlapping Cameras Tracking
Tracking Objects Across Multiple Moving Cameras
Multiple Camera Tracking
Our Approach
Second Layer (Track Generation)
Cross Camera Track Association
Camera-1 as a Constraint
Track Refinement Constraints
Experimental Results Multiple Cameras
Summary
Taught by
UCF CRCV