Psalm001/leaves-recognition

KWASU Computer Science Final Year Project Title: A Comparative Analysis of Convolutional Neural Network (CNN) and Decision Tree for the Recognition of Medicinal Plant Leaves 🌿 Medicinal Plant Leaf Recognition A machine learning and computer vision application for recognising medicinal plant species from leaf images. --- 📌 Overview Medicinal Plant Leaf Recognition is a machine learning project that explores the classification of medicinal plants using leaf images. The project compares two different machine learning approaches: 1. Decision Tree classification using extracted image features 2. CNN feature extraction combined with a Random Forest classifier The trained models are integrated into an interactive Streamlit web application, allowing users to upload a leaf image and obtain predictions from the different approaches. The application also provides information about the processed image and allows users to compare the prediction approaches and their processing performance. --- 🎯 Problem Statement Identifying medicinal plants manually can be difficult, particularly when different plant species have visually similar leaves. Computer vision and machine learning can assist with plant identification by learning visual patterns from leaf images. This project investigates how different feature-representation and classification strategies perform when applied to medicinal plant recognition. The main question explored by the project is: «How effectively can traditional machine-learning features and deep-learning features be used to recognise medicinal plant species from leaf images?» --- 💡 Project Approach The project implements two main recognition pipelines. Approach 1 — Decision Tree The first approach uses manually extracted image characteristics as input to a Decision Tree classifier. Leaf Image │ ▼ Image Preprocessing │ ▼ Feature Extraction │ ├── Colour Features ├── Texture Features └── Shape / Image Features │ ▼ Decision Tree │ ▼ Predicted Plant This approach demonstrates how traditional machine learning can be applied to image classificatio …