Lung nodule detection and segmentation using a patch-based multi-atlas method

Mustafa Alam, Ganesh Sankaranarayanan, Venkat Devarajan

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

CT image based lung nodule detection is the most widely used and accepted method for detecting lung cancer. Most CT image based methods are based on supervised/unsupervised learning, which has a high number of false positives and needs a large amount pre-segmented training samples. This problem can be solved, if a set of optimally small number of training samples can be created, where each sample has lung nodules of similar size and shape as the target image of the actual patient. Based on this hypothesis, we propose a novel patch-based multi-atlas method with three main steps: a) a small set of atlases is selected by comparing the target image with a larger set of atlas images using a size-shape based feature vector, b) lung nodules are selected using a patch-based method, where each pixel of a target image is labelled by comparing the image patch, centered by the pixel with patches from an atlas library and choosing the most probable labels according to a defined closest match criterion and c) Laplacian of Gaussian blob detection method is used to find the segmented area of the lung nodule. We tested the method with more than 5 test images, where each test image is applied to more than 200 atlas images. For non-attached nodules in the size between 3 mm to 30 mm, sensitivity of the proposed algorithm is 100%.

Original languageEnglish (US)
Title of host publicationProceedings - 2016 International Conference on Computational Science and Computational Intelligence, CSCI 2016
EditorsMary Yang, Hamid R. Arabnia, Leonidas Deligiannidis, Leonidas Deligiannidis
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages23-28
Number of pages6
ISBN (Electronic)9781509055104
DOIs
StatePublished - Mar 17 2017
Externally publishedYes
Event2016 International Conference on Computational Science and Computational Intelligence, CSCI 2016 - Las Vegas, United States
Duration: Dec 15 2016Dec 17 2016

Publication series

NameProceedings - 2016 International Conference on Computational Science and Computational Intelligence, CSCI 2016

Conference

Conference2016 International Conference on Computational Science and Computational Intelligence, CSCI 2016
Country/TerritoryUnited States
CityLas Vegas
Period12/15/1612/17/16

Keywords

  • Atlas based segmentation
  • Cancer Detection
  • CT
  • Medical image
  • Pulmonary nodule detection

ASJC Scopus subject areas

  • Computer Science Applications
  • Information Systems
  • Health Informatics
  • Artificial Intelligence
  • Computer Networks and Communications

Fingerprint

Dive into the research topics of 'Lung nodule detection and segmentation using a patch-based multi-atlas method'. Together they form a unique fingerprint.

Cite this