sihwa-park / nebla

NeBLa: Neural Beer-Lambert for 3D Reconstruction of Oral Structures from Panoramic Radiographs

Welcome to the official code repository for the paper titled "NeBLa: Neural Beer-Lambert for 3D Reconstruction of Oral Structures from Panoramic Radiographs," accepted at AAAI 2024. In this repository, you'll find the core model classes used in the generation module.

Model Details

The final encoder-decoder model for refinement is a 3D-UNet model, and the corresponding code can be found at pytorch-3dunet.

A significant portion of the MLP model and point embedder is adapted from the code available at nerf-pytorch.

For the translation module, the code is based on the CycleGAN model, and you can refer to pytorch-CycleGAN-and-pix2pix.

Feel free to explore and adapt the code to suit your needs.

nebla's Issues

Are there any plans to release the full code?