Development of a Web-Based Platform Utilizing a Hybrid CNN Architecture for Real-Time Fecal Image Classification

Abstract

Accurate classification of fecal images offers a non-invasive, scalable method for diagnosing enteric diseases such as Coccidiosis, Salmonellosis, and Newcastle disease, overcoming limitations associated with traditional diagnostic approaches that are often time-intensive, costly, and reliant on expert interpretation. This study presents the development of a real-time, web-based diagnostic platform employing a hybrid Convolutional Neural Network (CNN) model to classify fecal images. The proposed architecture integrates MobileNetV2 for efficient inference with VGG-16 for robust feature extraction, optimizing both diagnostic accuracy and computational performance. The model was trained on a curated and augmented dataset of labeled fecal images, subsequently quantized and converted into TensorFlow Lite format to ensure compatibility with mobile devices. The deployed system allows users to upload images through a web browser or Android application, delivering immediate classification results across four categories: Coccidiosis, Salmonella, Newcastle disease, or Healthy. Evaluation results demonstrate high diagnostic accuracy and responsiveness in real-world conditions. The findings highlight the potential of hybrid deep learning models within mobile and web-based platforms to facilitate field-ready, real-time disease detection. Future enhancements should include dataset expansion across diverse geographic regions and the integration of additional diagnostic modalities to improve robustness and adaptability.