Image Based Cattle Nose Print Recognition Technology:

Abstract

Cow identification is a critical aspect of livestock management, contributing to efficient record-keeping, insurance, and tracking systems. Traditional methods are labour-intensive and prone to errors. In this thesis, we propose an automated system for cow identification using nose print images. We evaluated several models, including Support Vector Machine (SVM), K-Nearest Neighbor (KNN), VGG16, VGG19, and a custom Convolutional Neural Network (CNN). Our custom CNN model demonstrated superior performance with an accuracy of 99.31% and classification scores (precision, recall, and F1-score) all at 0.99. This high level of accuracy indicates the potential of the proposed system to replace manual identification methods, thereby enhancing the reliability and efficiency of livestock management. Our findings suggest that the custom CNN model is particularly well-suited for this application, offering a robust solution for cow identification that can be integrated into existing livestock management systems in Bangladesh. This research highlights the significance of leveraging advanced machine learning techniques for improving agricultural practices and livestock management.