Plant Disease Diagnosis with Deep Learning & Explainable AI

Team project of four students: a deep learning model that aims, from a simple photo of a leaf, to identify the crop, the disease and its type (fungal, bacterial, viral…) with a confidence score, and to show which area of the image its answer is based on. The challenge is to work on real field photos, not only on lab images. My role: the model and its training.

Technologies: Python · PyTorch · timm · EfficientNet · Grad-CAM · Streamlit · Google Colab