Bananini is a web application that allows users to upload a photo of a banana leaf for disease diagnosis. It contains two pretrained models, which were finetuned to classify banana leaves into one of four categories: Cordana, Pestalotiopsis, Sigatoka, and Healthy. Each disease classification is supplemented with its corresponding characteristics and description.

Our first image classification model uses MobileNetV2, a lightweight Convolutional Neural Network (CNN). We removed its original classifier and added a custom head tailored to banana disease detection, including fully connected layers and dropout for regularization. Initially, the base layers were frozen to retain learned image features, and only the new head was trained. In a second phase, deeper layers were unfrozen and fine-tuned to improve performance.

Our second model leverages a Vision Transformer (ViT). We fine-tuned the pre-trained ViT model on our dataset, allowing it to adapt its global attention-based features for accurate classification.