Cited time in webofscience Cited time in scopus

3D Femur Feature Reconstruction with 2D X-ray Images for Surgical Navigation

Title
3D Femur Feature Reconstruction with 2D X-ray Images for Surgical Navigation
Author(s)
Jongho Chien
DGIST Authors
Jongho ChienJaesung HongSanghyun Joung
Advisor
홍재성
Co-Advisor(s)
Sanghyun Joung
Issued Date
2022
Awarded Date
2022/08
Type
Thesis
Subject
Statistical Shape Model, 3D Reconstruction, Femur Modeling
Description
Statistical Shape Model, 3D Reconstruction, Femur Modeling
Abstract
One of the main approaches to three-dimensional (3D) reconstruction of human anatomy which does not use computed tomography (CT) is the statistical shape model (SSM) method, which reconstructs a 3D model using an SSM and X-ray images. When we develop a patient-specific 3D reconstruction of a femur modeled using the SSM and X-ray images, the target shape should not vary outside the range of variations allowed by the SSM built from a training dataset. To cover more severe variations in the target shape, we propose the shape-partitioned statistical shape model (SPSSM). With this model, a shape can be divided into several segments of anatomical interest. The salient feature of this method is that an SSM for each segment can be described using a matrix obtained by partitioning the original eigenvector matrix, which consists of eigenvectors arranged as columns, without segmenting the shape and building an independent SSM for each segment. To quantify the reconstruction error of the proposed method, we generated two groups of deformation models of the femur which cannot be easily represented by the conventional SSM. One group of femurs had an anteversion angle deformation, and the other group of femurs had two different scales of the femoral head. Each experiment was performed using the leave-one-out method for twelve femurs. When the femoral head was rotated by 30˚, the average reconstruction error of the conventional SSM was 5.34 mm, which was reduced to 3.82 mm for the proposed SPSSM. When the femoral head size was decreased by 20%, the average reconstruction error of the SSM was 4.70 mm, which was reduced to 3.56 mm for the SPSSM. When the femoral head size was increased by 20%, the average reconstruction error of the SSM was 4.28 mm, which was reduced to 3.10 mm for the SPSSM. The experimental results for the two groups of deformation models showed that the proposed SPSSM outperformed the conventional SSM.|컴퓨터 단층 촬영(CT)을 사용하지 않는 인체 해부학의 3차원 재구성에 대한 주요 접근 방식 중 하나는 statistical shape model(SSM)과 2D X-ray images을 사용하여 3D 모델을 재구성하는 방법이다. 2D X-ray image를 사용하여 SSM을 환자 맞춤형 3D 모델로 재구성할 때, 재구성 되는 모형은 트레이닝 데이터세트로부터 결정되는 SSM이 나타낼 수 있는 변형 범위를 벗어날 수 없다. 본 논문에서는 평균에서 많이 벗어난 변화를 다루기 위해 shape-partitioned statistical shape model(SPSSM)을 제안한다. 제안하는 방법을 사용하면 형상을 해부학적으로 관심 있는 여러 부분으로 나누어 3D 모델 재구성을 진행하게 된다. 이 방법의 특징은 형상을 분할하고 분할된 각 부분에 대해 독립적인 SSM을 따로 구축하지 않고 트레이닝 데이터세트의 공분산 행렬로부터 만들어진 고유 백터로 구성된 고유 백터 행렬을 분할하여 얻은 행렬을 사용하여 각 부분에 대한 SSM을 다루는 점이다. 제안된 방법의 재구성 오차를 측정하기 위해 기존 SSM으로 쉽게 표현할 수 없는 두 가지 변형 대퇴골 모델을 생성하여 재구성 오차 측정 실험을 진행하였다. 한 그룹의 대퇴골은 anteversion 변형을 가지게 생성 되었고, 다른 그룹의 대퇴골은 대퇴골 머리 부분이 두 개의 서로 다른 크기를 가지도록 생성 되었다. 각 실험은 12개의 대퇴골에 대해 one-leave-out 방법을 사용하여 수행되었다. 대퇴골의 anteversion을 30도 회전시켰을 때, 기존의 SSM의 평균 재구성 오차는 5.34mm이었고, 제안된 SPSSM의 경우 3.82mm로 감소하였다. 대퇴골 머리의 크기를 20% 감소시켰을 때 평균 재구성 오차는 4.70mm에서 3.56mm로 감소하였다. 대퇴골의 머리의 크기를 20% 증가시켰을 때 평균 재구성 오차는 4.28mm에서 3.10mm로 감소하였다. 두 그룹의 변형 모델에 대한 실험 결과는 제안된 SPSSM이 기존의 SSM을 더 낳은 결과를 보여주었다.
Table Of Contents
I. Introduction 1
1.1 Surgical navigation and patient-specific 3D reconstruction 1
1.2 Statistical Shape Model (SSM) and 3D reconstruction with 2D images 2
1.3 Proposed method 4
II. Method 6
2.1 System workflow 6
2.2 SSM 7
2.3 Shape-partitioned statistical shape model 8
2.4 Femur Models 10
2.5 Partitioning model for SPSSM 10
2.6 3D/2D registration 12
2.7 3D/2D correspondence 17
2.8 Optimization 19
2.9 Effects of changing shape parameters on the femur shape 22
III. Experiments and results 24
3.1 Experiment on 3D/2D registration error tolerance and initial image alignment 24
3.2 Experiment on weighting parameters of optimization 26
3.3 Image database of X-ray images and DRRs 27
3.4 Deformation of femoral anteversion angle 28
3.5 Scale deformation of femoral head 42
IV. Discussion 48
V. Conclusion 51
URI
http://dgist.dcollection.net/common/orgView/200000628966

http://hdl.handle.net/20.500.11750/16787
DOI
10.22677/thesis.200000628966
Degree
Doctor
Department
Department of Robotics and Mechatronics Engineering
Publisher
DGIST
Related Researcher
  • 홍재성 Hong, Jaesung
  • Research Interests Surgical Navigation; Surgical Robot; Medical Imaging; 영상 유도 수술 로봇; 수술 내비게이션
Files in This Item:

There are no files associated with this item.

Appears in Collections:
Department of Robotics and Mechatronics Engineering Theses Ph.D.

qrcode

  • twitter
  • facebook
  • mendeley

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

BROWSE