AI 221 Final Project
Role
Technologies
Repository
View on GitHub→Overview
Accurate plant species identification is vital in botany, agriculture, and environmental research, but traditional methods are often slow and require expert input. This project tackles the task of classifying plant species from leaf images using computer vision and machine learning, with a primary focus on classical machine learning approaches that leverage extracted image features.
Implementation
The core methodology centers on extracting predefined features from leaf images—such as shape descriptors, texture analysis, and color histograms—and applying classical machine learning algorithms (including SVMs, Random Forests, and k-NN) to classify the species based on these features. For comparative benchmarking, we also evaluated deep learning models using pretrained convolutional neural networks (CNNs). Throughout the project, we explored various feature engineering and model selection strategies, ultimately demonstrating the effectiveness of classical algorithms for leaf-based plant identification, with deep learning models serving as a reference point for performance comparison.