Document Detail


Efficient brain lesion segmentation using multi-modality tissue-based feature selection and support vector machines.
MedLine Citation:
PMID:  23303595     Owner:  NLM     Status:  MEDLINE    
Abstract/OtherAbstract:
Support vector machines (SVM) are machine learning techniques that have been used for segmentation and classification of medical images, including segmentation of white matter hyper-intensities (WMH). Current approaches using SVM for WMH segmentation extract features from the brain and classify these followed by complex post-processing steps to remove false positives. The method presented in this paper combines advanced pre-processing, tissue-based feature selection and SVM classification to obtain efficient and accurate WMH segmentation. Features from 125 patients, generated from up to four MR modalities [T1-w, T2-w, proton-density and fluid attenuated inversion recovery(FLAIR)], differing neighbourhood sizes and the use of multi-scale features were compared. We found that although using all four modalities gave the best overall classification (average Dice scores of 0.54  ±  0.12, 0.72  ±  0.06 and 0.82  ±  0.06 respectively for small, moderate and severe lesion loads); this was not significantly different (p = 0.50) from using just T1-w and FLAIR sequences (Dice scores of 0.52  ±  0.13, 0.71  ±  0.08 and 0.81  ±  0.07). Furthermore, there was a negligible difference between using 5 × 5 × 5 and 3 × 3 × 3 features (p = 0.93). Finally, we show that careful consideration of features and pre-processing techniques not only saves storage space and computation time but also leads to more efficient classification, which outperforms the one based on all features with post-processing.
Authors:
Jean-Baptiste Fiot; Laurent D Cohen; Parnesh Raniga; Jurgen Fripp
Related Documents :
12067365 - Multi-field 3d scanning light microscopy of early embryogenesis.
23811635 - Procedure time and the determination of polypoid abnormalities with experience: impleme...
19617885 - Long-term, high-resolution imaging in the mouse neocortex through a chronic cranial win...
12508695 - Flow cytometry and factor analysis evaluation of confocal image sequences of morphologi...
Publication Detail:
Type:  Journal Article; Research Support, Non-U.S. Gov't     Date:  2013-01-10
Journal Detail:
Title:  International journal for numerical methods in biomedical engineering     Volume:  29     ISSN:  2040-7947     ISO Abbreviation:  Int J Numer Method Biomed Eng     Publication Date:  2013 Sep 
Date Detail:
Created Date:  2013-09-09     Completed Date:  2014-03-04     Revised Date:  2014-07-09    
Medline Journal Info:
Nlm Unique ID:  101530293     Medline TA:  Int J Numer Method Biomed Eng     Country:  England    
Other Details:
Languages:  eng     Pagination:  905-15     Citation Subset:  IM    
Copyright Information:
Copyright © 2013 John Wiley & Sons, Ltd.
Export Citation:
APA/MLA Format     Download EndNote     Download BibTex
MeSH Terms
Descriptor/Qualifier:
Brain / anatomy & histology,  pathology
Brain Neoplasms / pathology*
Humans
Image Processing, Computer-Assisted / methods*
Magnetic Resonance Imaging
Neuroimaging / methods*
Reproducibility of Results
Support Vector Machines*

From MEDLINE®/PubMed®, a database of the U.S. National Library of Medicine


Previous Document:  Uncovering selection bias in case-control studies using Bayesian post-stratification.
Next Document:  Polyoxometalates as a Novel Class of Artificial Proteases: Selective Hydrolysis of Lysozyme under Ph...