000203681 001__ 203681
000203681 005__ 20190117192126.0
000203681 022__ $$a0278-0062
000203681 02470 $$2ISI$$a000353899600008
000203681 0247_ $$2doi$$a10.1109/TMI.2014.2376274
000203681 037__ $$aARTICLE
000203681 245__ $$aLearning Structured Models for Segmentation of 2D and 3D Imagery
000203681 269__ $$a2015
000203681 260__ $$bInstitute of Electrical and Electronics Engineers$$c2015
000203681 336__ $$aJournal Articles
000203681 520__ $$aEfficient and accurate segmentation of cellular structures in microscopic data is an essential task in medical imaging. Many state-of-the-art approaches to image segmentation use structured models whose parameters must be carefully chosen for optimal performance. A popular choice is to learn them using a large-margin framework and more specifically structured support vector machines (SSVM). Although SSVMs are appealing, they suffer from certain limitations. First, they are restricted in practice to linear kernels because the more powerful non-linear kernels cause the learning to become prohibitively expensive. Second, they require iteratively finding the most violated constraints, which is often intractable for the loopy graphical models used in image segmentation. This requires approximation that can lead to reduced quality of learning. In this article, we propose three novel techniques to overcome these limitations. We first introduce a method to “kernelize” the features so that a linear SSVM framework can leverage the power of non-linear kernels without incurring much additional computational cost. Moreover, we employ a working set of constraints to increase the reliability of approximate subgradient methods and introduce a new way to select a suitable step size at each iteration. We demonstrate the strength of our approach on both 2D and 3D electron microscopic (EM) image data and show consistent performance improvement over state-of-the-art approaches.
000203681 6531_ $$aimage processing
000203681 6531_ $$acomputer vision
000203681 6531_ $$aelectron microscopy
000203681 6531_ $$aimage segmentation
000203681 6531_ $$akernel methods
000203681 6531_ $$amitochondria
000203681 6531_ $$astatistical machine learning
000203681 6531_ $$astructured prediction
000203681 6531_ $$asegmentation
000203681 6531_ $$asuperpixels
000203681 6531_ $$asupervoxels
000203681 700__ $$0242715$$aLucchi, Aurélien$$g185205
000203681 700__ $$aMárquez-Neila, Pablo
000203681 700__ $$aBecker, Carlos
000203681 700__ $$0244419$$aLi, Yunpeng$$g201504
000203681 700__ $$0242712$$aSmith, Kevin$$g163328
000203681 700__ $$0240043$$aKnott, Graham$$g159872
000203681 700__ $$0240252$$aFua, Pascal$$g112366
000203681 773__ $$j34$$k5$$q1096-1110$$tIEEE Transactions on Medical Imaging
000203681 8564_ $$s7269713$$uhttps://infoscience.epfl.ch/record/203681/files/paper.pdf$$yPreprint$$zPreprint
000203681 8564_ $$s172879$$uhttps://infoscience.epfl.ch/record/203681/files/supplementary.pdf$$yAppendix$$zAppendix
000203681 909C0 $$0252025$$pCIME$$xU10192
000203681 909C0 $$0252087$$pCVLAB$$xU10659
000203681 909CO $$ooai:infoscience.tind.io:203681$$pSB$$pIC$$particle
000203681 917Z8 $$x243370
000203681 917Z8 $$x243370
000203681 917Z8 $$x112366
000203681 917Z8 $$x112366
000203681 937__ $$aEPFL-ARTICLE-203681
000203681 973__ $$aEPFL$$rREVIEWED$$sPUBLISHED
000203681 980__ $$aARTICLE