What is bone suppression?
๐ป Try hovering the mouse over the image below and moving it horizontally!
Chest X-rays (CXRs) are commonly utilized as a low-dose modality for lung screening. Nonetheless, the efficacy of CXRs is somewhat impeded, given that approximately 75% of the lung area overlaps with bone, which in turn hampers the detection and diagnosis of diseases. As a remedial measure, bone suppression techniques have been introduced. The current dual-energy subtraction imaging technique in the clinic requires costly equipment and subjects being exposed to high radiation. To circumvent these issues, deep learning-based image generation algorithms have been proposed. However, existing methods fall short in terms of producing high-quality images and capturing texture details, particularly with pulmonary vessels. To address these issues, this paper proposes a new bone suppression framework, termed BS-Diff, that comprises a conditional diffusion model equipped with a U-Net architecture and a simple enhancement module to incorporate an autoencoder. Our proposed network cannot only generate soft tissue images with a high bone suppression rate but also possesses the capability to capture fine image details. Additionally, we compiled the largest dataset since 2010, including data from 120 patients with high-definition, high-resolution paired CXRs and soft tissue images collected by our affiliated hospital. Extensive experiments, comparative analyses, ablation studies, and clinical evaluations indicate that the proposed BS-Diff outperforms several bone-suppression models across multiple metrics.
๐ To the best of our knowledge, this is the pioneering study that harnesses
diffusion models for the generation of soft-tissue images from CXRs, thus addressing and overcoming the prevailing limitations of
DES.
๐ In our enhancement module, we introduce an innovative amalgamation of
varied loss functions, devised to more effectively encapsulate
spatial features and intricate texture details of images, while concurrently preserving their overall structures.
๐ We have also assembled the most extensive dataset since 2010, comprising
high-resolution paired images from 120 patients, which were collected in collaboration with our partner hospital. We anticipate making this dataset publicly available in the near future.
๐ Through comprehensive experiments, comparative analyses, ablation studies, and clinical evaluations, we substantiate the
superior performance of our proposed
BS-Diff model in comparison to several high-performing bone suppression models.
Visualization of ablation studies on BS-Diff with and without enhancement module. 
@inproceedings{chen2024bs,
title={Bs-diff: Effective bone suppression using conditional diffusion models from chest x-ray images},
author={Chen, Zhanghao and Sun, Yifei and Ge, Ruiquan and Qin, Wenjian and Pan, Cheng and Deng, Wenming and Liu, Zhou and Min, Wenwen and Elazab, Ahmed and Wan, Xiang and others},
booktitle={2024 IEEE International Symposium on Biomedical Imaging (ISBI)},
pages={1--5},
year={2024},
organization={IEEE}
}