Animal ReID and Tracking



Author:   Tao Jiang
Date:   Feb 19, 2025



Abstract

In this project, we propose a framework for continuous 24/7 tracking and identification of cows on a farm. Using video footage captured by mounted cameras, the system automatically detects cows in each frame, recognizes their behaviors, identifies individuals through their freeze brands, assigns the corresponding freeze brand ID to each cow, and tracks their movement across frames. This framework enables real-time monitoring of individual cows and facilitates the analysis of behavioral patterns and health conditions, providing valuable insights for farm management and promoting improved animal welfare.






Implementation Details

Operating System: Windows
Programming Language: Python
Framework: PyTorch
IDE: Pycharm
GPU: Nvidia GeForce RTX 5060