3D Reconstruction of Leg Bones from X-Ray Images using CNN-based Feature Analysis

Hangkee Kim, Kisuk Lee, Dongchun Lee, Nakhoon Baek

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

19 Scopus citations

Abstract

3D reconstruction of subject specific bones from X-Ray images is an important issue in a variety of medical applications such as diagnosis. We propose a method to reconstruct 3D leg bones from 2D X-Ray images using feature analysis. First, bounding boxes are detected by Convolutional Neural Network(CNN). Feature points and feature ellipses are extracted from them. Such features are aligned with feature information of 3D bone model. Then, the boundary of X-Ray is detected from aligned boundary of 3D model. The 3D model is fine-tuned by adjusting the Statistical Shape Model parameter. We believe that this method makes 3D modeling easier by improving the automatic detection of feature information compared to the manual landmark input method.

Original languageEnglish
Title of host publicationICTC 2019 - 10th International Conference on ICT Convergence
Subtitle of host publicationICT Convergence Leading the Autonomous Future
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages669-672
Number of pages4
ISBN (Electronic)9781728108926
DOIs
StatePublished - Oct 2019
Event10th International Conference on Information and Communication Technology Convergence, ICTC 2019 - Jeju Island, Korea, Republic of
Duration: 16 Oct 201918 Oct 2019

Publication series

NameICTC 2019 - 10th International Conference on ICT Convergence: ICT Convergence Leading the Autonomous Future

Conference

Conference10th International Conference on Information and Communication Technology Convergence, ICTC 2019
Country/TerritoryKorea, Republic of
CityJeju Island
Period16/10/1918/10/19

Keywords

  • 3D reconstruction
  • CNN
  • feature analysis
  • leg bone
  • SSM
  • X-Ray

Fingerprint

Dive into the research topics of '3D Reconstruction of Leg Bones from X-Ray Images using CNN-based Feature Analysis'. Together they form a unique fingerprint.

Cite this