A Divide-and-Conquer Approach towards Understanding Deep Networks

Fu W, Breininger K, Schaffert R, Ravikumar N, Maier A (2019)

Publication Language: English

Publication Type: Conference contribution, Original article

Publication year: 2019

Publisher: Springer, Cham

City/Town: Shenzhen

Pages Range: 183-192

Conference Proceedings Title: Medical Image Computing and Computer Assisted Intervention (MICCAI 2019)

Event location: Shenzhen, China

ISBN: 978-3-030-32238-0

URI: https://link.springer.com/chapter/10.1007/978-3-030-32239-7_21

DOI: 10.1007/978-3-030-32239-7_21

Open Access Link: https://link.springer.com/chapter/10.1007/978-3-030-32239-7_21

Abstract

Deep neural networks have achieved tremendous success in various fields including medical image segmentation. However, they have long been criticized for being a black-box, in that interpretation, understanding and correcting architectures is difficult as there is no general theory for deep neural network design. Previously, precision learning was proposed to fuse deep architectures and traditional approaches. Deep networks constructed in this way benefit from the original known operator, have fewer parameters, and improved interpretability. However, they do not yield state-of-the-art performance in all applications. In this paper, we propose to analyze deep networks using known operators, by adopting a divide-and-conquer strategy to replace network components, whilst retaining networks performance. The task of retinal vessel segmentation is investigated for this purpose. We start with a high-performance U-Net and show by step-by-step conversion that we are able to divide the network into modules of known operators. The results indicate that a combination of a trainable guided filter and a trainable version of the Frangi filter yields a performance at the level of U-Net (AUC 0.974 vs. 0.972) with a tremendous reduction in parameters (111, 536 vs. 9, 575). In addition, the trained layers can be mapped back into their original algorithmic interpretation and analyzed using standard tools of signal processing.

Authors with CRIS profile

Related research project(s)

How to cite

APA:

Fu, W., Breininger, K., Schaffert, R., Ravikumar, N., & Maier, A. (2019). A Divide-and-Conquer Approach towards Understanding Deep Networks. In Dinggang Shen, Tianming Liu, Terry M. Peters, Lawrence H. Staib, Caroline Essert, Sean Zhou, Pew-Thian Yap, Ali Khan (Eds.), Medical Image Computing and Computer Assisted Intervention (MICCAI 2019) (pp. 183-192). Shenzhen, China, CN: Shenzhen: Springer, Cham.

MLA:

Fu, Weilin, et al. "A Divide-and-Conquer Approach towards Understanding Deep Networks." Proceedings of the MICCAI 2019, Shenzhen, China Ed. Dinggang Shen, Tianming Liu, Terry M. Peters, Lawrence H. Staib, Caroline Essert, Sean Zhou, Pew-Thian Yap, Ali Khan, Shenzhen: Springer, Cham, 2019. 183-192.

BibTeX: Download