2D Hand Pose Estimation from RGB Image

1. Description

Repository includes working PyTorch pipeline to train a model and make inferences.

Here are my tutorials that explain it in details:

• Gentle introduction to 2D Hand Pose Estimation: Approach Explained
• Gentle introduction to 2D Hand Pose Estimation: Let's Code It!

For this project, open-source FreiHAND dataset is used.

2. Repository Content

Repository has the following files:

• notebooks/Train Notebook.ipynb (notebook to train a model)
• notebooks/Inference Notebook.ipynb (notebook for inferences)
• utils/model.py (model class)
• utils/dataset.py (dataset class to load FreiHAND dataset)
• utils/trainer.py (class to train model)
• utils/prep_utils.py (supplementary functions)
• weights/model_final (trained model weights)
• requirements.txt (required packages & versions)

3. License

This project is licensed under the terms of the MIT license.