한국정보처리학회:학술대회논문집 (Annual Conference of KIPS)

한국정보처리학회 (Korea Information Processing Society)
권호 표지

Mask2Former 를 이용한 CT 및 PET 영상의 정밀 폐암 분할

Precision Lung Cancer Segmentation from CT & PET Images Using Mask2Former

  • ;
  • 김경백 (인공지능융합학과, 전남대학교)
  • Md Ilias Bappi (Dept. of. Artificial Intelligence Convergence, Chonnam National University) ;
  • Kyungbeak Kim (Dept. of. Artificial Intelligence Convergence, Chonnam National University)
  • 발행 : 2024.10.31
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

Lung cancer is a leading cause of death worldwide, highlighting the critical need for early diagnosis. Lung image analysis and segmentation are essential steps in this process, but manual segmentation of medical images is extremely time-consuming for radiation oncologists. The complexity of this task is heightened by the significant variability in lung tumors, which can differ greatly in size, shape, and texture due to factors like tumor subtype, stage, and patient-specific characteristics. Traditional segmentation methods often struggle to accurately capture this diversity. To address these challenges, we propose a lung cancer diagnosis system based on Mask2Former, utilizing CT (Computed Tomography) and PET (Positron Emission Tomography) images. This system excels in generating high-quality instance segmentation masks, enabling it to better adapt to the heterogeneous nature of lung tumors compared to traditional methods. Additionally, our system classifies the segmented output as either benign or malignant, leveraging a self-supervised network. The proposed approach offers a powerful tool for early diagnosis and effective management of lung cancer using CT and PET data. Extensive experiments demonstrate its effectiveness in achieving improved segmentation and classification results.

키워드

과제정보

This work was supported by Institute of Information & communications Technology Planning & Evaluation (IITP) under the Artificial Intelligence Convergence Innovation Human Resources Development (IITP-2023-RS-2023-00256629) grant funded by the Korea government(MSIT). This research was supported by the MSIT(Ministry of Science and ICT), Korea, under the ITRC(Information Technology Research Center) support program(IITP-2024-RS-2024-00437718) supervised by the IITP(Institute for Information & Communications Technology Planning & Evaluation).

참고문헌

  1. https://www.cancer.org/research/cancer-factsstatistics/all-cancer-facts-figures/2023-cancer-facts-figures.html
  2. Esteva, A., Kuprel, B., Novoa, R. A., et al. (2017). Dermatologist-level classification of skin cancer with deep neural networks. Nature, 542(7639), 115-118.
  3. Liu, Y., Shen, D., & Wang, L. (2019). Lung nodule detection and segmentation using deep learning. Medical Image Analysis, 58, 101569.
  4. Dosovitskiy, A., Beyer, L., & Kolesnikov, A. (2020). Image segmentation with Vision Transformers. IEEE Transactions on Pattern Analysis and Machine Intelligence, 42(6), 1301-1310.
  5. Chen, L., Li, X., & Wang, X. (2022). Mask2Former: A novel transformer-based model for instance segmentation. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2022, 2251-2260.
  6. Kim, Y., Park, S., & Lee, J. (2023). Multi-modal deep learning for enhanced lung cancer diagnosis using CT and PET scans. Medical Imaging: Analysis and Applications, 45, 159-172.
  7. Siegel, R.L.; Miller, K.D.; Fuchs, H.E.; Jemal, A. Cancer statistics. Cancer J. Clin. 2022, 72, 7-33.
  8. Li, P., Wang, S., Li, T., Lu, J., HuangFu, Y., & Wang, D. (2020). A Large-Scale CT and PET/CT Dataset for Lung Cancer Diagnosis (Lung-PET-CT-Dx) [Data set]. The Cancer Imaging Archive. https://doi.org/10.7937/TCIA.2020.NNC2-0461
  9. Stamatis, G.; Eberhard, W.; Pottgen, C. Surgery after multimodality treatment for non-small-cell lung cancer. Lung Cancer 2004, 45, S107-S112.
  10. Chiang, T.A.; Chen, P.H.; Wu, P.F.; Wang, T.N.; Chang, P.Y.; Ko, A.M.; Huang, M.S.; Ko, Y.C. Important prognostic factors for the long-term survival of lung cancer subjects in Taiwan. BMC Cancer 2008, 8, 324.
  11. Paing, M.P.; Hamamoto, K.; Tungjitkusolmun, S.; Visitsattapongse, S.; Pintavirooj, C. Automatic detection of pulmonary nodules using three-dimensional chain coding and optimized random forest. Appl. Sci. 2020, 10, 2346.
  12. Perez, G.; Arbelaez, P. Automated lung cancer diagnosis using three-dimensional convolutional neural networks. Med Biol. Eng. Comput. 2020, 58, 1803-1815.
  13. Kasinathan, G.; Jayakumar, S.; Gandomi, A.H.; Ramachandran, M.; Fong, S.J.; Patan, R. Automated 3-D lung tumor detection and classification by an active contour model and CNN classifier. Expert Syst. Appl. 2019, 134, 112-119.