Techniques in Machine Learning & Artificial Intelligence (C7082)

Dan Peters

The Cow Behavior Detection Model Project aims to develop a robust YOLOv8-based deep learning model to identify and classify cow behaviors (e.g., grazing, lying, standing and estrus) using bounding boxes in images. The project is designed to enhance livestock monitoring by automating behavior analysis, providing farmers and researchers with valuable insights into herd health and activity.

📂 Repository Contents

  • Cow_Behaviour_Detection_Model The model will download a selection of images from google drive and annotate named bounding boxes on the images in reference to behaviour
  • Directory: Key outputs, including images, videos, figures, tables, and other results generated by the notebook through th etraining process.

📑 Notebook Sections

(these are also available via the Gihub repo)

Development of Cow Behaviour Detection Model

(https://rpubs.com/DanJPeters/Development_of_Cow_Behaviour_Detection_Model)

Here, you’ll find: - Background - Methodology - A typical day working on the model - Discussion - Literature

Timeline of Cow Behaviour Detection Model

(https://rpubs.com/DanJPeters/Timeline_of_Cow_Behaviour_Detection_Model)

This section provides: - A step by step guide on how the development of how the model progressed over time - Information on techniques and experimentation

Results and Analysis of Cow Behaviour Detection Model

(https://rpubs.com/DanJPeters/Analysis_of_Cow_Behaviour_Detection_Model)

Explore: - Exploratory Data Analysis (EDA) with graphical outputs. - Principal diagnostics and key results from the applied methods.

You can find additional test images to download here if required:

(https://github.com/DMFP13/Cow-Behaviour-Detection/tree/main/cow_behavior_detection/test)