Semi-supervised peripapillary atrophy segmentation with shape constraint

Mengxuan Li; Weihang Zhang; Ruixiao Yang; Jie Xu; He Zhao; Huiqi Li

doi:10.1016/j.compbiomed.2023.107464

Semi-supervised peripapillary atrophy segmentation with shape constraint

Comput Biol Med. 2023 Sep 7:166:107464. doi: 10.1016/j.compbiomed.2023.107464. Online ahead of print.

Authors

Mengxuan Li¹, Weihang Zhang¹, Ruixiao Yang¹, Jie Xu², He Zhao³, Huiqi Li⁴

Affiliations

¹ Beijing Institute of Technology, Beijing 100081, China.
² Beijing Institute of Ophthalmology, Beijing Tongren Hospital, Capital Medical University, Beijing 100005, China.
³ Beijing Institute of Technology, Beijing 100081, China. Electronic address: zhaohe@bit.edu.cn.
⁴ Beijing Institute of Technology, Beijing 100081, China. Electronic address: huiqili@bit.edu.cn.

PMID: 37734355
DOI: 10.1016/j.compbiomed.2023.107464

Abstract

Peripapillary atrophy (PPA) is a clinical abnormality related to many eye diseases, such as myopia and glaucoma. The shape and area of PPA are essential indicators of disease progression. PPA segmentation is a challenging task due to blurry edge and limited labeled data. In this paper, we propose a novel semi-supervised PPA segmentation method enhanced by prior knowledge. In order to learn shape information in the network, a novel shape constraint module is proposed to restrict the PPA appearance based on active shape model. To further leverage large amount of unlabeled data, a Siamese-like model updated by exponential moving average is introduced to provide pseudo labels. The pseudo labels are further refined by region connectivity correction. Extensive experiments on a clinical dataset demonstrate that our proposed PPA segmentation method provides good qualitative and quantitative performance.

Keywords: Active shape model; Mean teacher model; Peripapillary atrophy segmentation; Semi-supervised; Shape constraint.