Learning features for tissue classification with the classification restricted Boltzmann machine

Gijs van Tulder*, Marleen de Bruijne

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

24 Citations (Scopus)

Abstract

Performance of automated tissue classification in medical imaging depends on the choice of descriptive features. In this paper, we show how restricted Boltzmann machines (RBMs) can be used to learn features that are especially suited for texture-based tissue classification. We introduce the convolutional classification RBM, a combination of the existing convolutional RBM and classification RBM, and use it for discriminative feature learning. We evaluate the classification accuracy of convolutional and non-convolutional classification RBMs on two lung CT problems. We find that RBM-learned features outperform conventional RBM-based feature learning, which is unsupervised and uses only a generative learning objective, as well as often-used filter banks. We show that a mixture of generative and discriminative learning can produce filters that give a higher classification accuracy.

Original languageEnglish
Title of host publicationMedical Computer Vision: Algorithms for Big Data : International Workshop, MCV 2014, Held in Conjunction with MICCAI 2014, Cambridge, MA, USA, September 18, 2014, Revised Selected Papers
EditorsBjoern Menze, Georg Langs, Albert Montillo, Michael Kelm, Henning Müller, Shaoting Zhang, Weidong (Tom) Cai, Dimitris Metaxas
Number of pages12
PublisherSpringer
Publication date2014
Pages47-58
Chapter5
ISBN (Print)978-3-319-13971-5
ISBN (Electronic)978-3-319-13972-2
DOIs
Publication statusPublished - 2014
EventInternational Workshop on Medical Computer Vision 2014 - Cambridge, United States
Duration: 18 Sep 201418 Sep 2014

Conference

ConferenceInternational Workshop on Medical Computer Vision 2014
Country/TerritoryUnited States
CityCambridge
Period18/09/201418/09/2014
SeriesLecture notes in computer science
ISSN0302-9743

Cite this