Shrimp Disease Classification and Medicine Recommendation using Deep Learning Techniques

Aquaculture sustainability is threatened by bacterial, viral, and fungal diseases that cause large financial losses in shrimp farming. Due to the high mortality rates caused by rapid disease spread, early and precise detection is necessary. Automated solutions are required because traditional manual inspection is laborious and prone to mistakes. DenseNet is used to detect and classify shrimp diseases, guaranteeing effective feature propagation, decreased redundancy, and increased accuracy of 98.89%. To improve model robustness, ImageDataGenerator is used for pre-processing and data augmentation. F1-score, recall, accuracy, and precision are used to assess performance. In addition to classifying diseases using DenseNet, the system recommends specific medications mapped to each identified disease through a rule-based lookup table curated from aquaculture literature, enabling timely and evidence-backed treatment suggestions. Our method increases diagnostic efficiency and classification accuracy by utilizing deep learning and transfer learning. According to experimental results, shrimp diseases can be identified with high effectiveness. Our system provides a quicker, more accurate, and more dependable way to detect shrimp disease than conventional techniques and other deep learning models like VGG16 (visual geometry group), Inception, and CNN (convolutional neural network) based architectures.