signlang-ml-project

Machine Learning-based Sign Language Detection System using HOG features and classical ML models.

This project implements a sign language detection system using Python. It uses classical machine learning algorithms (SVM, Random Forest, KNN) along with HOG feature extraction to recognize hand gestures from images. The system is trained on the Sign Language MNIST dataset and supports real-time prediction using a webcam.

Project structure includes:

• data/: Dataset files
• notebooks/: Jupyter notebooks for experimentation
• src/: Python scripts (training, preprocessing, inference)
• models/: Saved ML models and preprocessing artifacts
• signlang_env.yml: Conda environment for reproducibility