The main goal of this project is to predict patient''s pose using deep learning, and integrating it into our Unity3D pipeline and Swift XR application. To achieve this, I employed 2D CNN and PointNet-based architectures. Models are further integrated into our solution using Barracuda, Onnxruntime and CoreML.

• WebsiteRepository 1
• WebsiteRepository 2
• WebsiteUnity Onnxruntime Integration Unity Barracuda Integration
• PlatformUbuntu / Unity / CoreML
• StackPython, Tensorflow, Keras, Trimesh, Colab, PyTorch