Scalable, incremental learning with MapReduce parallelization for cell detection in high-resolution 3D microscopy data

Handle URI:
http://hdl.handle.net/10754/599558
Title:
Scalable, incremental learning with MapReduce parallelization for cell detection in high-resolution 3D microscopy data
Authors:
Sung, Chul; Woo, Jongwook; Goodman, Matthew; Huffman, Todd; Choe, Yoonsuck
Abstract:
Accurate estimation of neuronal count and distribution is central to the understanding of the organization and layout of cortical maps in the brain, and changes in the cell population induced by brain disorders. High-throughput 3D microscopy techniques such as Knife-Edge Scanning Microscopy (KESM) are enabling whole-brain survey of neuronal distributions. Data from such techniques pose serious challenges to quantitative analysis due to the massive, growing, and sparsely labeled nature of the data. In this paper, we present a scalable, incremental learning algorithm for cell body detection that can address these issues. Our algorithm is computationally efficient (linear mapping, non-iterative) and does not require retraining (unlike gradient-based approaches) or retention of old raw data (unlike instance-based learning). We tested our algorithm on our rat brain Nissl data set, showing superior performance compared to an artificial neural network-based benchmark, and also demonstrated robust performance in a scenario where the data set is rapidly growing in size. Our algorithm is also highly parallelizable due to its incremental nature, and we demonstrated this empirically using a MapReduce-based implementation of the algorithm. We expect our scalable, incremental learning approach to be widely applicable to medical imaging domains where there is a constant flux of new data. © 2013 IEEE.
Citation:
Sung C, Woo J, Goodman M, Huffman T, Choe Y (2013) Scalable, incremental learning with MapReduce parallelization for cell detection in high-resolution 3D microscopy data. The 2013 International Joint Conference on Neural Networks (IJCNN). Available: http://dx.doi.org/10.1109/IJCNN.2013.6706769.
Publisher:
Institute of Electrical and Electronics Engineers (IEEE)
Journal:
The 2013 International Joint Conference on Neural Networks (IJCNN)
KAUST Grant Number:
KUSC1-016-04
Issue Date:
Aug-2013
DOI:
10.1109/IJCNN.2013.6706769
Type:
Conference Paper
Sponsors:
This publication is based in part on work supported by Award No. KUSC1-016-04, made by King Abdullah University of Science and Technology(KAUST); NIH/NINDS grant #R01-NS54252, NSF grants #0079874,#0905041, and #1208174; and by the Breakout Labs. Computing time onAmazon EC2 was generously donated by Amazon. Chul Sung was supportedby 3Scan as an intern while conducting part of this research. We would liketo thank Bruce H. McCormick, Bernard Mesa, Louise C. Abbott, DavidMayerich, and Jaerock Kwon for their contributions in data acquisition, andSugato Basu for helpful comments on machine learning and MapReduce.
Appears in Collections:
Publications Acknowledging KAUST Support

Full metadata record

DC FieldValue Language
dc.contributor.authorSung, Chulen
dc.contributor.authorWoo, Jongwooken
dc.contributor.authorGoodman, Matthewen
dc.contributor.authorHuffman, Todden
dc.contributor.authorChoe, Yoonsucken
dc.date.accessioned2016-02-28T05:53:20Zen
dc.date.available2016-02-28T05:53:20Zen
dc.date.issued2013-08en
dc.identifier.citationSung C, Woo J, Goodman M, Huffman T, Choe Y (2013) Scalable, incremental learning with MapReduce parallelization for cell detection in high-resolution 3D microscopy data. The 2013 International Joint Conference on Neural Networks (IJCNN). Available: http://dx.doi.org/10.1109/IJCNN.2013.6706769.en
dc.identifier.doi10.1109/IJCNN.2013.6706769en
dc.identifier.urihttp://hdl.handle.net/10754/599558en
dc.description.abstractAccurate estimation of neuronal count and distribution is central to the understanding of the organization and layout of cortical maps in the brain, and changes in the cell population induced by brain disorders. High-throughput 3D microscopy techniques such as Knife-Edge Scanning Microscopy (KESM) are enabling whole-brain survey of neuronal distributions. Data from such techniques pose serious challenges to quantitative analysis due to the massive, growing, and sparsely labeled nature of the data. In this paper, we present a scalable, incremental learning algorithm for cell body detection that can address these issues. Our algorithm is computationally efficient (linear mapping, non-iterative) and does not require retraining (unlike gradient-based approaches) or retention of old raw data (unlike instance-based learning). We tested our algorithm on our rat brain Nissl data set, showing superior performance compared to an artificial neural network-based benchmark, and also demonstrated robust performance in a scenario where the data set is rapidly growing in size. Our algorithm is also highly parallelizable due to its incremental nature, and we demonstrated this empirically using a MapReduce-based implementation of the algorithm. We expect our scalable, incremental learning approach to be widely applicable to medical imaging domains where there is a constant flux of new data. © 2013 IEEE.en
dc.description.sponsorshipThis publication is based in part on work supported by Award No. KUSC1-016-04, made by King Abdullah University of Science and Technology(KAUST); NIH/NINDS grant #R01-NS54252, NSF grants #0079874,#0905041, and #1208174; and by the Breakout Labs. Computing time onAmazon EC2 was generously donated by Amazon. Chul Sung was supportedby 3Scan as an intern while conducting part of this research. We would liketo thank Bruce H. McCormick, Bernard Mesa, Louise C. Abbott, DavidMayerich, and Jaerock Kwon for their contributions in data acquisition, andSugato Basu for helpful comments on machine learning and MapReduce.en
dc.publisherInstitute of Electrical and Electronics Engineers (IEEE)en
dc.titleScalable, incremental learning with MapReduce parallelization for cell detection in high-resolution 3D microscopy dataen
dc.typeConference Paperen
dc.identifier.journalThe 2013 International Joint Conference on Neural Networks (IJCNN)en
dc.contributor.institutionTexas A and M University, College Station, United Statesen
dc.contributor.institutionCalifornia State University Los Angeles, Los Angeles, United Statesen
dc.contributor.institution3Scan, San Francisco, CA, United Statesen
kaust.grant.numberKUSC1-016-04en
All Items in KAUST are protected by copyright, with all rights reserved, unless otherwise indicated.