matt-pettit/signspotter

KNN Optimization for ASL Recognition 👋

This project aims to design and implement a machine learning system capable of accurately classifying individual hand signs from static images representing the 24 letters of the American Sign Language (ASL) alphabet, excluding i and j. The system utilizes the K-Nearest Neighbor (KNN) algorithm and explores different hyperparameters to optimize its performance. The dataset is from kaggle: https://www.kaggle.com/datasets/ash2703/handsignimages

Authors 💻

• Holly Lewis
• Matt Pettit

Project Overview 📖

The project investigates the effectiveness of the KNN algorithm in classifying static images of ASL letters. It explores the impact of different distance metrics (Euclidean, Manhattan, and Minkowski) and K values on the model's accuracy, precision, recall, F1 score, and Jaccard index. Key features:

• 🖼️ Classification of 24 ASL letters from static images
• 🔍 Evaluation of different distance metrics and K values
• 📈 Performance analysis using various metrics (accuracy, precision, recall, F1 score, Jaccard index)
• 📊 Visualization of results with confusion matrices and plots

License 📄

This project is licensed under the MIT License.