Improvement of subtle microcalcifications detection in DBT slices

Eun Joon Kim, Dae Hoe Kim, Eun Suk Cha, Yong Man Ro

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

Abstract

In this paper, a novel microcalcification detection in Digital Breast Tomosynthesis (DBT) is proposed, which allows detecting subtle microcalcifications that have relatively low contrast in DBT slices. 3D Foveal segmentation is employed for volume of interest (VOI) detection using adaptive thresholding scheme by local volume contrast. Furthermore, in order to conserve structural characteristics of clustered microcalcifications, features from VOI maximum ray-tracing images with three orthogonal directions are extracted. Extensive and comparative experiments were carried out to demonstrate the effectiveness of the proposed 3D Foveal segmentation and the feature extraction method on clinical DBT cases. The results showed that the proposed microcalcification detection achieves a sensitivity of 98% at 0.5 FP per volume.

Original languageEnglish
Title of host publication2014 IEEE-EMBS International Conference on Biomedical and Health Informatics, BHI 2014
PublisherIEEE Computer Society
Pages322-325
Number of pages4
ISBN (Print)9781479921317
DOIs
StatePublished - 2014
Event2014 IEEE-EMBS International Conference on Biomedical and Health Informatics, BHI 2014 - Valencia, Spain
Duration: 1 Jun 20144 Jun 2014

Publication series

Name2014 IEEE-EMBS International Conference on Biomedical and Health Informatics, BHI 2014

Conference

Conference2014 IEEE-EMBS International Conference on Biomedical and Health Informatics, BHI 2014
Country/TerritorySpain
CityValencia
Period1/06/144/06/14

Fingerprint

Dive into the research topics of 'Improvement of subtle microcalcifications detection in DBT slices'. Together they form a unique fingerprint.

Cite this