Livestock Behavior Recognition using TinyML for Real-Time Edge-Based Monitoring

This project presents a lightweight livestock behavior recognition system using Tiny Machine Learning (TinyML) and computer vision techniques. The system is designed to monitor cattle behavior in real time using a camera feed without relying on cloud infrastructure or additional hardware sensors. A compact Convolutional Neural Network (CNN) model is trained to classify behaviors such as grazing, fence proximity, and no-animal conditions. The trained model is converted to TensorFlow Lite format and deployed for edge-based inference. The system generates alerts when risky behavior such as fence proximity is detected. The objective of this project is to provide a low-cost, energy-efficient, and practical solution for smart livestock monitoring in rural farming environments.