Content-based retrieval of brain diffusion magnetic resonance image

Siqi Liu*, Nur Hadi, Sidong Liu, Sonia Pujol, Ron Kikinis, Fan Zhang, Dagan Feng, Weidong Cai

*Corresponding author for this work

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

1 Citation (Scopus)


The content-based retrieval of diffusion magnetic resonance (dMR) imaging data would enable a wide range of analyses on large databases with dMR images.This paper proposes a content-based retrieval framework for dMR images to explore the use of Diffusion Tensor Imaging (DTI) - derived parameters. The propagation graph algorithm is proposed for the query-centric retrieval of dMR subjects and the fusion of different features. The proposed framework was evaluated with ADNI database with 233 baseline dMR images. The preliminary results show that the proposed retrieval framework is able to retrieve subjects with similar neurodegenerative patterns.

Original languageEnglish
Title of host publicationMultimodal Retrieval in the Medical Domain
Subtitle of host publicationFirst International Workshop, MRMD 2015, Revised Selected Papers
EditorsHenning Müller H., Oscar Alfonso Jimenez del Toro, Allan Hanbury, Georg Langs, Antonio Foncubierta Rodríguez
Place of PublicationCham
PublisherSpringer-VDI-Verlag GmbH & Co. KG
Number of pages7
ISBN (Electronic)9783319244716
ISBN (Print)9783319244709
Publication statusPublished - 1 Jan 2015
Externally publishedYes
Event1st International Workshop on Multimodal Retrieval in the Medical Domain, MRMD 2015 - Vienna, Austria
Duration: 29 Mar 201529 Mar 2015

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Conference1st International Workshop on Multimodal Retrieval in the Medical Domain, MRMD 2015

Fingerprint Dive into the research topics of 'Content-based retrieval of brain diffusion magnetic resonance image'. Together they form a unique fingerprint.

Cite this