Automating Poultry Disease Detection using Deep Learning

Abstract

Poultry farming plays a vital role in global food production but the emerging threat of diseases poses significant challenges to both sustainability and food security. In particular, this research study investigates the integration of deep learning techniques to automate the detection of four major poultry diseases—Avian Influenza, Coccidiosis, Newcastle Disease, and Gumboro Disease—from faecal samples. The proposed methodology involves collecting diverse faecal samples, pre-processing the data, and developing a Convolutional Neural Network (CNN) architecture. The CNN layered architecture is designed to extract hierarchical features and learn complex patterns associated with each disease. Through the integration of activation function, Rectified Linear Units (ReLU), the network incorporates non-linearity, enhancing its ability to detect the disease-related features. The faecal samples undergo image enhancement, normalization, and segmentation to ensure suitability for the deep learning model. The performance of the proposed model is evaluated using the performance metrics and achieved an overall accuracy of 98.82% on the training set, 93.22% on the testing set, and 96.65% on the validation set., precision, recall and F1-Score. This research study contributes to the advancement of automated disease detection, offering a potential solution to mitigate the impact of poultry diseases and enhance overall food safety.