wlgjs8 / Descriptor-based-Registration

Implementation of "CBCT-Dental Scan Registration via Descriptor Representation Learning"

CBCT-Dental Scan Registration via Descriptor Representation Learning

Implementation of "CBCT-Dental Scan Registration via Descriptor Representation Learning"

Framework

Representation Learning Guidance via CT gradient

Prerequisites

• Python 3.6.6
• Pytorch 1.9.1

Installation

The required packages are located in requirements.txt.

pip install torch==1.9.1+cu111 torchvision==0.10.1+cu111 torchaudio==0.9.1 -f https://download.pytorch.org/whl/torch_stable.html

pip install -r requirements.txt

Training

To train our detection baseline, follow below train instruction.

python train/train_0_detection.py 
python train/train_1_detection_and_metal_classification.py 

Testing

To test the performance of our method, place the given pretrained weight inside of checkpoints/detection_best_model.pth. To test the performance of our method, place the given pretrained weight inside of checkpoints/detection_and_metal_classification_best_model.pth.

python test/test_0_detection.py
python test/test_1_detection_and_metal_classification.py