Cited time in webofscience Cited time in scopus

Full metadata record

DC Field Value Language
dc.contributor.advisor 박석호 -
dc.contributor.author Yun-Jeong Lee -
dc.date.accessioned 2024-02-29T21:01:42Z -
dc.date.available 2024-02-29T21:01:42Z -
dc.date.issued 2024 -
dc.identifier.uri http://hdl.handle.net/20.500.11750/48071 -
dc.identifier.uri http://dgist.dcollection.net/common/orgView/200000726122 -
dc.description Image-free tumor segmentation;Palpation;Surface Reconstruction;Stiffness Estimation;RMIS -
dc.description.abstract In tumor resection surgery, one of the essential processes that helps doctors make appropriate diagnosis and surgery for patients is tumor segmentation. A common method to perform tumor segmentation is to use medical imaging and palpation together. However, because this method is subjective, it has the limitation that there may be differences between doctors. As one of the methods to solve this problem, automated tumor segmentation is being used. However, automated tumor segmentation has limitations due to errors in medical imaging itself and lack of quantitative tactile information. Therefore, in this paper, an image-free tumor segmentation method using robotic palpation in MIS is proposed to solve these limitations.
In this study, the proposed method adopted a process called Surface Reconstruction, which can obtain 3D surface information of the tissue. This process includes Point Cloud Acquisition and Point2Mesh Deep learning algorithm. The accuracy of the Surface Reconstruction process was verified by analyzing the Root Mean Square Error (RMSE) through phantom experiments. Through this verification, it was confirmed that the accuracy was less than 1.1 mm. In addition, to obtain tissue stiffness information, the stiffness estimation based on the reconstructed surface was used. Especially, for viscoelastic tissues such as organs, tissue stiffness data was obtained by the indentation method after a probe path for accurate stiffness estimation was created. The obtained data were visualized as a stiffness map. Finally, the proposed image-free tumor segmentation was verified in terms of specificity and sensitivity through phantom and ex-vivo experiments.

Keywords: Image-Free tumor segmentation, Palpation, Surface Reconstruction, Stiffness Estimation, RMIS
|Tumor resection surgery 에서, 의사가 환자의 적합한 진단 및 수술을 할 수 있도록 도와주는 필수적인 과정 중 하나는 tumor segmentation 이다. 이러한 Tumor segmentation 을 수행하기 위한 일반적인 방법으로는 medical imaging 과 palpation 을 함께 사용하는 것이 있다. 하지만 이 방법은 주관적이기 때문에 의사들 간의 차이가 있을 수 있다는 한계가 있다. 이를 해결하기 위한 방법 중 하나로, automated tumor segmentation 이 활용되고 있다. 그러나 automated tumor segmentation 은 medical imaging 자체의 error 와 정량적 촉각 정보 부족의 한계점을 가진다. 그러므로, 본 논문에서는 이러한 한계들을 해결하기 위해 MIS에서 robotic palpation을 사용하여 image-free 한 tumor segmentation 방법이 제안된다.
본 연구에서, 제안된 방법에는 조직의 3D surface information 을 얻을 수 있는 Surface Reconstruction 이라는 과정이 채택되었다. 이 과정은 Point Cloud Acquisition 과 Point2Mesh Deep learning algorithm 을 포함한다. Surface Reconstruction 과정의 정확성 검증은 phantom 실험을 통해 Root Mean Square Error(RMSE)를 분석하여 수행되었다. 이 검증을 통해, 1.1 mm 이하의 정확도를 보임이 확인되었다. 또한, 조직의 강성 정보를 알기 위해, Reconstructed surface 를 기반으로 한 Stiffness Estimation 이 사용되었다. 특히, 장기와 같은 viscoelastic 한 조직에 대해, 정확한 강성 추정을 위한 Probe Path 가 생성된 후 indentation 방법에 의해 조직의 stiffness 데이터가 얻어졌다. 얻어진 데이터는 stiffness map 으로 가시화되었다. 마지막으로, phantom 실험과 Ex-vivo 실험을 통해, specificity 와 sensitivity 의 관점에서 제안된 Image-free 한 종양 세분화가 검증되었다.
-
dc.description.tableofcontents Ⅰ. Introduction 1
1.1 Tumor segmentation 1
1.2 Limitations of tumor segmentation 1
1.3 Proposal 2
1.4 Structure of Thesis 4

ⅠⅠ. Method & Experiment 5
2.1 Surface Reconstruction 5
2.1.1 Point cloud acquisition using Teleoperation 5
2.1.2 Point2Mesh Deep Learning Algorithm 12
2.2 Stiffness Estimation 14
2.2.1 Probe Path Generation 14
2.2.2 Indentation Method 19
2.3 Tumor Segmentation 21
2.4 Phantom Experiment 24
2.4.1 Experimental Setup 24
2.4.2 Experimental Procedure 30
2.5 Ex-vivo Experiment 31
2.5.1 Experimental Setup & Procedure 31

ⅠⅠⅠ. Result & Discussion 33
3.1 Phantom Experiment 33
3.2 Ex-vivo Experiment 42
3.3 Discussion 44

ⅠV. Conclusion 47
4.1 Conclusion 47

V. Appendix 50
5.1 Safety of tissue contact 50
-
dc.format.extent 52 -
dc.language eng -
dc.publisher DGIST -
dc.title Image-free Tumor Segmentation of Soft Tissue using Minimally Invasive Robotic Palpation -
dc.type Thesis -
dc.identifier.doi 10.22677/THESIS.200000726122 -
dc.description.degree Master -
dc.contributor.department Department of Robotics and Mechatronics Engineering -
dc.contributor.coadvisor Jaesung Hong -
dc.date.awarded 2024-02-01 -
dc.publisher.location Daegu -
dc.description.database dCollection -
dc.citation XT.RM이66 202402 -
dc.date.accepted 2024-01-30 -
dc.contributor.alternativeDepartment 로봇및기계전자공학과 -
dc.subject.keyword Image-free tumor segmentation -
dc.subject.keyword Palpation -
dc.subject.keyword Surface Reconstruction -
dc.subject.keyword Stiffness Estimation -
dc.subject.keyword RMIS -
dc.contributor.affiliatedAuthor Yun-Jeong Lee -
dc.contributor.affiliatedAuthor Sukho Park -
dc.contributor.affiliatedAuthor Jaesung Hong -
dc.contributor.alternativeName 이윤정 -
dc.contributor.alternativeName Sukho Park -
dc.contributor.alternativeName 홍재성 -
dc.rights.embargoReleaseDate 2026-02-28 -
Files in This Item:

There are no files associated with this item.

Appears in Collections:
Department of Robotics and Mechatronics Engineering Theses Master

qrcode

  • twitter
  • facebook
  • mendeley

Items in Repository are protected by copyright, with all rights reserved, unless otherwise indicated.

BROWSE