Using Deep Learning for Facial Recognition: Implementing a Convolutional Siamese Neural Network in a Facial Recognition System for Cattle

David McDaid, Kevin Meehan

Abstract: Technology is changing the way humans live each day and is also helping transform the area of Agriculture. Sensors, Cameras, and Drones have significantly reduced the physical labor required in Agriculture. There is an opportunity to create a scalable facial recognition solution that will be able to correctly identify individual cattle in a herd. This will allow farmers to effectively manage data regarding their herd. This research aims to evaluate the effectiveness of applying a Convolutional Siamese Neural Network to this problem area. A bespoke dataset containing 32 individual cows was created and evaluated in this research. The Siamese Network was evaluated against validation and test data, the model produced an accuracy result of 90%.