Curvature-of-Cervical-Spine-Estimation

This repository includes the official implementation of the paper:

HTN: Hybrid Transformer Network for Curvature of Cervical Spine Estimation

(State: Review in Second Round, On Applied Sciences, MDPI)

Authors and affiliations: Yifan Yao 1, 2, Jiuqing Dong 2, Wenjun Yu 1, and Yongbin Gao 1,* 1 International Joint Research Lab of Intelligent Perception and Control, Shanghai University of Engineering Science, No. 333 Longteng Road, Shanghai 201620, P.R China 2 Division of Computer Science and Engineering, Jeonbuk National University, Jeonju 54896, Korea

The code and our model will be released in two weeks.

Install

Clone repo and install related library in a Python>=3.6.0 environment, including PyTorch>=1.5.

Training

python main.py --data_dir dataPath --num_epoch 100 --batch_size 2 --dataset spinal --phase train

Testing and visualization

python main.py --resume weightPath --data_dir dataPath --dataset spinal --phase test

Evaluation and result scatter plot

python main.py --resume weightPath --data_dir dataPath --dataset spinal --phase eval

Note

if you have any problems, please contact us.