Plant Disease Detection (Final Year Project, A*)

This final-year project (graded A*) is a full-stack web application that intakes plant images, performs object detection to isolate leaves, removes visual noise, and classifies each isolated image before presenting a summary of the plant's health.

The frontend was prototyped in Figma and built in React, with a Django backend for storing and displaying detection history, and a CNN trained with TensorFlow/Keras for the core classification task.

• Developed a full-stack web app performing object detection, noise removal, and disease classification on plant images.
• Prototyped the interface in Figma and implemented it with React for seamless disease detection.
• Built a secure Django backend API for storing and displaying detection result history with a SQL database.
• Developed and trained a Convolutional Neural Network (CNN) using TensorFlow and Keras for plant disease detection.
• Created a labelled dataset using CVAT and applied transfer learning to boost model performance.
• Evaluated and improved the model by comparing performance metrics and tuning hyperparameters.

Detection & Classification Pipeline

Object detection to isolate leaves, noise removal, and CNN-based classification, summarized in a plant health report.