A New Ensemble Learning Framework for 3D Biomedical Image Segmentation

  • Hao Zheng University of Notre Dame
  • Yizhe Zhang University of Notre Dame
  • Lin Yang University of Notre Dame
  • Peixian Liang University of Notre Dame
  • Zhuo Zhao University of Notre Dame
  • Chaoli Wang University of Notre Dame
  • Danny Z. Chen University of Notre Dame

Abstract

3D image segmentation plays an important role in biomedical image analysis. Many 2D and 3D deep learning models have achieved state-of-the-art segmentation performance on 3D biomedical image datasets. Yet, 2D and 3D models have their own strengths and weaknesses, and by unifying them together, one may be able to achieve more accurate results. In this paper, we propose a new ensemble learning framework for 3D biomedical image segmentation that combines the merits of 2D and 3D models. First, we develop a fully convolutional network based meta-learner to learn how to improve the results from 2D and 3D models (base-learners). Then, to minimize over-fitting for our sophisticated meta-learner, we devise a new training method that uses the results of the baselearners as multiple versions of “ground truths”. Furthermore, since our new meta-learner training scheme does not depend on manual annotation, it can utilize abundant unlabeled 3D image data to further improve the model. Extensive experiments on two public datasets (the HVSMR 2016 Challenge dataset and the mouse piriform cortex dataset) show that our approach is effective under fully-supervised, semisupervised, and transductive settings, and attains superior performance over state-of-the-art image segmentation methods.

Published
2019-07-17
Section
AAAI Technical Track: Machine Learning