Deep Learning Based Silkworm Disease Prediction and Medicine Suggestion System

The research exhibits an automated framework capable of identifying major silkworm diseases using a hybrid deep-learning architecture that combines a support to Convolutional Neural Network (CNN) with a technology-based EfficientNet-B0, VGG-16, ResNet-50 and MobileNet-V2 model. The EfficientNet-B0 network, pretrained on ImageNet, enhances feature extraction efficiency, while the CNN strengthens domain-specific learning to identify the accuracy level. Silkworm images are classified into four types: flacherie, grasserie, muscardine, and pebrine. Experimental evaluations indicate the CNN model EfficientNet-B0 approach achieves an accuracy of approximately 99% and classifies strong generalisation capabilities in real-time agriculture environments. The findings highlight the effectiveness of the AI-driven approach in disease monitoring and the potential to improve the yield by identifying the disease at early stages. There are two types of decision-making: healthy and unhealthy. Silkworm quality will be good by identifying the disease and industry development.