X-ray 영상에서 VHS와 콥 각도 자동 추출을 위한흉추 분할 기법
A Thoracic Spine Segmentation Technique for Automatic Extraction of VHS and Cobb Angle from X-ray Images
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
본 논문에서는 X-ray 영상에서 의료 진단지표를 자동으로 추출하기 위한 조직분할 기법을 제안한다. 척추질환이나 심장질환에 대한 진단지표로서, 흉추-심장 비율이나 콥 각도 등의 지표를 산출하기 위해서는 흉부 X-ray 영상으로부터 흉추, 용골 및 심장의 영역을 정확하게 분할하는 과정이필요하다. 본 연구에서는 이를 위하여 계층별로 영상의 고해상도의 표현과 저해상도의 특징지도로 변환되는 구조가 병렬적으로 연결되는 형태의심층신경망 모델을 채택하였다. 이러한 구조는 영상에서 세부 조직의 상대적인 위치정보가 분할 과정에 효과적으로 반영될 수 있게 한다. 또한픽셀 정보와 객체 정보가 다단계의 과정으로 상호 작용되는 OCR 모듈과, 네트워크의 각 채널이 서로 다른 가중치 값으로 반영되도록 하는 채널어텐션 모듈을 결합하여 학습 성능을 개선할 수 있음을 보인다. 부수적으로 X-ray 영상에서 피사체의 위치 변화, 형태의 변형 및 크기 변이 등에도강인한 성능을 제공하기 위하여 학습데이터를 증강하는 방법을 제시하였다. 총 145개의 인체 흉부 X-ray 영상과, 총 118개의 동물 X-ray 영상을사용한 실험을 통하여 제안된 이론의 타당성을 평가하였다.
Abstract
In this paper, we propose an organ segmentation technique for the automatic extraction of medical diagnostic indicators from X-rayimages. In order to calculate diagnostic indicators of heart disease and spinal disease such as VHS(vertebral heart scale) and Cobb angle,it is necessary to accurately segment the thoracic spine, carina, and heart in a chest X-ray image. A deep neural network model in whichthe high-resolution representation of the image for each layer and the structure converted into a low-resolution feature map are connectedin parallel was adopted. This structure enables the relative position information in the image to be effectively reflected in the segmentationprocess. It is shown that learning performance can be improved by combining the OCR module, in which pixel information and objectinformation are mutually interacted in a multi-step process, and the channel attention module, which allows each channel of the networkto be reflected as different weight values. In addition, a method of augmenting learning data is presented in order to provide robustperformance against changes in the position, shape, and size of the subject in the X-ray image. The effectiveness of the proposed theorywas evaluated through an experiment using 145 human chest X-ray images and 118 animal X-ray images.
참고문헌 (15)
- J. Yu, "FFANet : Feature fusion attention network to medical image segmentation", Biomedical Signal Processing and Control, 69, pp.102~912 (2021)
- Y. Hong, "Deep dual-resolution networks for real-time and accurate semantic segmentation of road scenes", arXiv preprint arXiv : 2101. 06085, ~ (2021)
- H. Yan, "Lawin transformer : Improving semantic segmentation transformer with multiscale representations via large window attention", arXiv preprint arXiv : 2201. 01615, ~ (2022)
- O. Ronneberger, "U-Net : Convolutional networks for biomedical image segmentation", International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI), pp.234~241 (2015)
- H. Wang, "Max-deeplab : End-to-end panoptic segmentation with mask transformers", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp.5463~5474 (2021)
- S. Bhattacharya, "ClamNet : Using contrastive learning with variable depth Unets for medical image segmentation", arXiv preprint arXiv : 2206. 05225, ~ (2022)
- Y. Chen, "VertXNet : Automatic segmentation and identification of lumbar and cervical vertebrae from spinal X-ray images", arXiv preprint arXiv : 2207. 05476, ~ (2022)
- N. Tajbakhsh, "Embracing imperfect datasets : A review of deep learning solutions for medical image segmentation", Medical Image Analysis, 63, pp.101~693 (2020)
- J. Jiang, "Multiple resolution residually connected feature streams for automatic lung tumor segmentation for CT images", IEEE Transactions on Medical Imaging, 38(1), pp.134~144 (2019)
- K. Sun, "High-resolution representations for labeling pixels and regions", arXiv preprint arXiv : 1904. 04514, ~ (2019)
- Y. Lee, "Thoracic spine segmentation of X-ray images using a modified HRNet", Proceedings of the Annual Spring Conference of Korea Information Processing Society Conference (KIPS) 2022, 29, pp.705~707 (2022)
- S. Han, "A deep learning technique to automatically extract VHS from X-ray images", Proceedings of Korea Computer Congress 2022 (KCC 2022), 29, pp.2075~2077 (2022)
- Y. Yuan, "Object-contextual representations for semantic segmentation", European Conference on Computer Vision (ECCV), pp.173~190 (2020)
- J. Hu, "Squeeze-and-excitation networks", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp.7132~7141 (2018)
- X. Wang, "ChestX-Ray8 : Hospital-scale chest X-ray database and benchmarks on weakly-supervised classification and localization of common thorax diseases", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp.2097~2106 (2017)