Weakly Supervised Deep Learning-based Methods for Brain Image Analysis

79.1K

views

59

authors

11

articles

Editorial

11 November 2022

Editorial: Weakly supervised deep learning-based methods for brain image analysis

Hancan Zhu

,

Mingxia Liu

,

Zhenyu Tang

and

Shuai Wang

5,711 views

0 citations

Editors

4

Shaoxing University

University of North Carolina at Chapel Hill

Beihang University

Shandong University

Impact

About

Deep learning-based methods have achieved excellent performance in various fields of brain image analysis. Most of the existing deep learning-based methods usually rely on large-scale datasets with high-quality full annotations. However, to acquire such data is usually time-consuming and requires rich expert experience. Moreover, because of Individual differences in experience and understanding, large-scale and full annotated datasets may suffer from large intra- and inter-observer variability, which could hinder their application in brain image analysis. In contrast, weak yet low-cost annotations (such as coarse annotations, partial annotations, or small sample annotations) are far easier to collect than high-quality full detailed annotations. Consequently, there is a strong desire for innovative deep learning-based methodologies that can efficiently learn from weak annotation data and achieve competitive performance compared with using full annotation data.

Deep learning-based methods have achieved excellent performance in various fields of brain image analysis. Most of the existing deep learning-based methods usually rely on large-scale datasets with high-quality full annotations. However, to acquire such data is usually time-consuming and requires rich expert experience. Moreover, because of Individual differences in experience and understanding, large-scale and full annotated datasets may suffer from large intra- and inter-observer variability, which could hinder their application in brain image analysis. In contrast, weak yet low-cost annotations (such as coarse annotations, partial annotations, or small sample annotations) are far easier to collect than high-quality full detailed annotations. Consequently, there is a strong desire for innovative deep learning-based methodologies that can efficiently learn from weak annotation data and achieve competitive performance compared with using full annotation data.

Share

Editors

Shaoxing University

University of North Carolina at Chapel Hill

Beihang University

Shandong University

Impact

79,109 Total views

65,131 Article views

12,347 Article downloads

1,631 Topic views

Published In

Journal Thumbnail

Frontiers in Neuroinformatics

About Frontiers Research Topics

With their unique mixes of varied contributions from Original Research to Review Articles, Research Topics unify the most influential researchers, the latest key findings and historical advances in a hot research area! Find out more on how to host your own Frontiers Research Topic or contribute to one as an author.

Suggest a topic