Machine Learning

CNN - Image Classification

Plant disease detection via Convolutional Neural Networks with Transfer Learning on MobileNetV2 - benchmarked across multiple architectures.

Python Keras MobileNetV2 Scikit-learn NumPy Matplotlib Google Colab
CNN Plant Classifier
The Problem

Plant diseases cause significant crop losses globally. Early, automated detection using computer vision can help farmers act quickly. The challenge is evaluating which deep learning approaches work best given the practical constraints of small dataset sizes and limited compute resources.

The Solution

A systematic comparison of three architectures — baseline CNN, deeper custom CNN, and fine-tuned MobileNetV2 — using Transfer Learning to leverage ImageNet feature detectors. Fine-tuning the upper layers adapts the model to plant disease patterns while preserving low-level feature knowledge.

Key Features
—Multiple CNN architectures built and benchmarked from scratch
—Transfer Learning using pretrained MobileNetV2 weights
—Fine-tuning strategy to adapt ImageNet features to plant disease domain
—Data augmentation pipeline to improve generalisation
—Evaluation with confusion matrices and classification reports
—Visualisations of training curves and feature maps
← Previous ThreeSixty Next → Network Intrusion Detection