qgking / MM-CL

code for "Multi-modality contrastive learning for sarcopenia screening from hip X-rays and clinical information" in MICCAI 2023

MM-CL

Multi-modality contrastive learning for sarcopenia screening from hip X-rays and clinical information published in MICCAI 2023

by Qiangguo Jin, Changjiang Zou, Changming Sun, Hui Cui, Changming Sun, Shu-Wei Huang, Yi-Jie Kuo, Ping Xuan, Leilei Cao, Ran Su, Leyi Wei, Henry B.L. Duh, and Yu-Pin Chen

Example results

• The framework of our proposed MM-CL. MM-CL is composed of (a) Non-local CAM Enhancement, (b) Visual-text Feature Fusion, and (c) Auxiliary contrastive representation..

• Visual interpretation of high-level features using t-SNE. The red and blue circles are sarcopenia and non-sarcopenia instances respectively.

• Sarcopenia diagnosis performance of recently proposed methods.

Citation

If the code is helpful for your research, please consider citing:

  @inproceedings{JIN2023MMCL,
      title = "Multi-modality contrastive learning for sarcopenia screening from hip X-rays and clinical information",
      booktitle = "International Conference on Medical Image Computing and Computer-Assisted Intervention",
      year = "2023",
      organization = "Springer",
      author = "Qiangguo Jin, Changjiang Zou, Changming Sun, Hui Cui, Changming Sun, Shu-Wei Huang, Yi-Jie Kuo, Ping Xuan, Leilei Cao, Ran Su, Leyi Wei, Henry B.L. Duh, and Yu-Pin Chen",
  }

Questions

General questions, please contact '[e-mail]'