Influence of decoder size for binary segmentation tasks in medical imaging

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

2 Citations (Scopus)

Abstract

Symmetric design of encoder-decoder networks is common in deep learning. For almost all segmentation problems, the output segmentation is vastly less complex compared to the input image. However, the effect of the size of the decoder on segmentation performance has not been investigated in literature. This work investigates the effect of reducing decoder size on binary segmentation performance in a medical imaging application. To this end, we propose a methodology to reduce the size of the decoder in encoder-decoder networks, where residual skip connections are employed in combination with a 1×1 convolution instead of concatenations (as employed by U-Net) to achieve models with asymmetric design. The results on the ISIC2017 data set show that the amount of trainable parameters in the decoder can be reduced by up to a factor 100 compared to standard U-Net, while retaining segmentation performance. Additionally, the reduced amount of trainable decoder parameters in the proposed models leads to inference times up to 3 times faster compared to standard U-Net.

Original languageEnglish
Title of host publicationMedical Imaging 2020
Subtitle of host publicationImage Processing
EditorsIvana Isgum, Bennett A. Landman
PublisherSPIE
ISBN (Electronic)9781510633933
DOIs
Publication statusPublished - 2020
EventSPIE Medical Imaging 2020 - Houston, United States
Duration: 15 Feb 202020 Feb 2020

Publication series

NameProceedings of SPIE
Volume11313
ISSN (Print)1605-7422

Conference

ConferenceSPIE Medical Imaging 2020
Country/TerritoryUnited States
CityHouston
Period15/02/2020/02/20

Keywords

  • Binary segmentation
  • Encoder-decoder networks
  • Model optimization
  • Skin lesions

Fingerprint

Dive into the research topics of 'Influence of decoder size for binary segmentation tasks in medical imaging'. Together they form a unique fingerprint.

Cite this