Non-Intrusive Detection of Poultry Diseases through Faecal Analysis using Deep Learning

This research conducts a comprehensive comparative study on the application of advanced deep learning models - Baseline Convolutional Neural Network, EfficientNetB3, DenseNet-121, AlexNet, and ResNet50 - for the early detection of diseases in broiler chickens using faecal image analysis. By integrating a detailed examination of each model’s performance, the study not only identifies the most effective techniques for non-intrusive health monitoring but also explores the integration of sophisticated data preprocessing and augmentation strategies. Utilizing You Only Look Once version 3 (YOLOv3) for precise region extraction, the study aims to enhance poultry health management by offering a nuanced understanding of how these models can be optimized for agricultural disease detection. The research’s findings and methodologies have the potential to influence future academic research and practical implementations in this field, paving the way for more efficient and effective disease management strategies in poultry farming and related agricultural sectors. This advancement could play a crucial role in enhancing the overall quality and safety of poultry products, benefiting consumers and the agricultural industry alike.