Automated Segmentation of Multiple Sclerosis Lesions Using Multi-dimensional Gated Recurrent Units

被引:46
作者
Andermatt, Simon [1 ]
Pezold, Simon [1 ]
Cattin, Philippe C. [1 ]
机构
[1] Univ Basel, Dept Biomed Engn, Allschwil, Switzerland
来源
BRAINLESION: GLIOMA, MULTIPLE SCLEROSIS, STROKE AND TRAUMATIC BRAIN INJURIES, BRAINLES 2017 | 2018年 / 10670卷
关键词
MD-GRU; MDGRU; Automatic MS lesion segmentation;
D O I
10.1007/978-3-319-75238-9_3
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
We analyze the performance of multi-dimensional gated recurrent units on automated lesion segmentation in multiple sclerosis. The segmentation of these pathologic structures is not trivial, since location, shape and size can be arbitrary. Furthermore, the inherent class imbalance of about 1 lesion voxel to 10 000 healthy voxels further exacerbates the correct segmentation. We introduce a new MD-GRU setup, using established techniques from the deep learning community as well as our own adaptations. We evaluate these modifications by comparing them to a standard MD-GRU network. We demonstrate that using data augmentation, selective sampling, residual learning and/or DropConnect on the RNN state can produce better segmentation results. Reaching rank #1 in the ISBI 2015 longitudinal multiple sclerosis lesion segmentation challenge, we show that a setup which combines these techniques can outperform the state of the art in automated lesion segmentation.
引用
收藏
页码:31 / 42
页数:12
相关论文
共 12 条
[1]   Multi-dimensional Gated Recurrent Units for the Segmentation of Biomedical 3D-Data [J].
Andermatt, Simon ;
Pezold, Simon ;
Cattin, Philippe .
DEEP LEARNING AND DATA LABELING FOR MEDICAL APPLICATIONS, 2016, 10008 :142-151
[2]  
[Anonymous], 2015, ARXIV150203167CS
[3]  
[Anonymous], 2014, ARXIV14092329CS
[4]   Longitudinal multiple sclerosis lesion segmentation: Resource and challenge [J].
Carass, Aaron ;
Roy, Snehashis ;
Jog, Amod ;
Cuzzocreo, Jennifer L. ;
Magrath, Elizabeth ;
Gherman, Adrian ;
Button, Julia ;
Nguyen, James ;
Prados, Ferran ;
Sudre, Carole H. ;
Cardoso, Manuel Jorge ;
Cawley, Niamh ;
Ciccarelli, Olga ;
Wheeler-Kingshott, Claudia A. M. ;
Ourselin, Sebastien ;
Catanese, Laurence ;
Deshpande, Hrishikesh ;
Maurel, Pierre ;
Commowick, Olivier ;
Barillot, Christian ;
Tomas-Fernandez, Xavier ;
Warfield, Simon K. ;
Vaidya, Suthirth ;
Chunduru, Abhijith ;
Muthuganapathy, Ramanathan ;
Krishnamurthi, Ganapathy ;
Jesson, Andrew ;
Arbel, Tal ;
Maier, Oskar ;
Handeles, Heinz ;
Iheme, Leonardo O. ;
Unay, Devrim ;
Jain, Saurabh ;
Sima, Diana M. ;
Smeets, Dirk ;
Ghafoorian, Mohsen ;
Platel, Bram ;
Birenbaum, Ariel ;
Greenspan, Hayit ;
Bazin, Pierre-Louis ;
Calabresi, Peter A. ;
Crainiceanu, Ciprian M. ;
Ellingsen, Lotta M. ;
Reich, Daniel S. ;
Prince, Jerry L. ;
Pham, Dzung L. .
NEUROIMAGE, 2017, 148 :77-102
[5]  
Cho K., 2014, P 2014 C EMP METH NA, P1724
[6]  
Cooijmans T., 2016, ARXIV160309025CS
[7]   Intra- and inter-observer agreement of brain MRI lesion volume measurements in multiple sclerosis - A comparison of techniques [J].
Filippi, M ;
Horsfield, MA ;
Bressi, S ;
Martinelli, V ;
Baratti, C ;
Reganati, P ;
Campi, A ;
Miller, DH ;
Comi, G .
BRAIN, 1995, 118 :1593-1600
[8]  
He K., 2015, PROC IEEE C COMPUT V
[9]  
Liao Q., 2016, ARXIV160403640CS
[10]  
Marijn F., 2015, P 28 INF C NEUR INF, P2998