High performance computing techniques for scaling image analysis workflows

Patrick M. Widener, Tahsin Kurc, Wenjin Chen, Fusheng Wang, Lin Yang, Jun Hu, Vijay Kumar, Vicky Chu, Lee Cooper, Jun Kong, Ashish Sharma, Tony Pan, Joel H. Saltz, David J. Foran

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

1 Scopus citations

Abstract

Biomedical images are intrinsically complex with each domain and modality often requiring specialized knowledge to accurately render diagnosis and plan treatment. A general software framework that provides access to high-performance resources can make possible high-throughput investigations of micro-scale features as well as algorithm design, development and evaluation. In this paper we describe the requirements and challenges of supporting microscopy analyses of large datasets of high-resolution biomedical images. We present high-performance computing approaches for storage and retrieval of image data, image processing, and management of analysis results for additional explorations. Lastly, we describe issues surrounding the use of high performance computing for scaling image analysis workflows.

Original languageEnglish
Title of host publicationApplied Parallel and Scientific Computing - 10th International Conference, PARA 2010, Revised Selected Papers
Pages67-77
Number of pages11
EditionPART 2
DOIs
StatePublished - 2012
Externally publishedYes
Event10th International Conference on Applied Parallel and Scientific Computing, PARA 2010 - Reykjavik, Iceland
Duration: Jun 6 2010Jun 9 2010

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 2
Volume7134 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference10th International Conference on Applied Parallel and Scientific Computing, PARA 2010
Country/TerritoryIceland
CityReykjavik
Period06/6/1006/9/10

Fingerprint

Dive into the research topics of 'High performance computing techniques for scaling image analysis workflows'. Together they form a unique fingerprint.

Cite this