torchreid

Local, API-free multi-stage identity pipeline: face recognition → liveness detection → super-resolution fallback → person Re-ID, built for realistic CCTV-style conditions.

Person Re-Identification pipeline (YOLOv8 + TorchReID/OSNet) — the fallback stage for matching people by appearance when face recognition can't resolve an identity.

This project implements a person detection and tracking system using YOLOv8 for real-time object detection, Deep SORT for object tracking, and OSNet for person re-identification. The model assigns unique IDs to each person and tracks them throughout the video, even after occlusion or re-entry into the frame.

Successfully developed a real-time soccer player re-identification and tracking system using YOLOv11 and Deep SORT with TorchreID-based embeddings.

A six-phase mechanistic audit of shortcut learning in a standard ResNet50 re-ID baseline trained on Market-1501.