Advanced Poultry Disease Detection with Deep Convolutional Neural Networks

Coccidiosis, Salmonella, and Newcastle represent prevalent poultry diseases that have the potential to significantly diminish chicken productivity if not identified promptly. In Tanzania, the insufficient accessibility of agricultural support services for chicken farmers hampers the timely recognition of diseases. The utilization of deep learning algorithms demonstrates potential in enabling early detection of specific poultry diseases. This study developed a deep Convolutional Neural Network (CNN) model to aid in the detection of chicken diseases by categorizing fecal pictures as healthy or unhealthy. Unhealthy fecal pictures can indicate Coccidiosis, Salmonella, and Newcastle infections. In the initial phase, the deep depth decision method is developed with optimal specifications using a deep convolutional neural network (DNN). The model is then formulated, trained on a dataset, and validated. The categorization is performed using a deep CNN, which achieves an accuracy of 91.47% and a corresponding training loss of around 0.2296. This study contributes to ongoing efforts to modernize poultry illness diagnosis using deep learning techniques. The model provided here offers a promising technique to detecting coccidiosis quickly and precisely, ultimately contributing to improved chicken health, increased productivity, and the promotion of sustainable poultry farming practices.