A VR-based Automated Strabismus Diagnosis System with Progressive Semi-Supervised Learning

Dehui Qiu; Bowei Ma; Ze Xiong; Yuhao Wang; Liguo Deng; Longfei Zhou; Xiaojie Cao; Weiwei Chen; Yumeng Wang; Meixia Jiang; Jing Fu; Fa Zhang; Xiaohua Wan

doi:10.1109/JBHI.2026.3679532

A VR-based Automated Strabismus Diagnosis System with Progressive Semi-Supervised Learning

Dehui Qiu^*
, Bowei Ma
, Ze Xiong
, Yuhao Wang
, Liguo Deng
, Longfei Zhou
, Xiaojie Cao
, Weiwei Chen
, Yumeng Wang
, Meixia Jiang
, Jing Fu
, Fa Zhang
, Xiaohua Wan

^*此作品的通讯作者

Capital Normal University
Capital Medical University
Beijing Institute of Technology

科研成果: 期刊稿件 › 文章 › 同行评审

摘要

Strabismus is a prevalent ocular disorder that can impair visual development and cause psychological issues if not diagnosed early. Conventional clinical diagnosis primarily relies on the prism cover test (PCT), which is subjective, requires patient cooperation, and lacks standardization. Recent advances in virtual reality (VR) and deep learning offer promising solutions for automated and standardized diagnosis. However, practical deployment faces three key challenges: realistic VR simulation of clinical exams, addressing image degradation (reflections/occlusions) with limited annotated data, and precise quantification of ocular deviations. In this study, we propose a novel VR-based automated strabismus diagnosis system by leveraging semi-supervised deep learning, and introduce a new clinical dataset, TongRenD. The framework incorporates five standardized clinical examination scenarios within a VR environment to ensure diagnostic consistency. We introduce ProgNet: an uncertainty-guided progressive semi-supervised segmentation network that integrates a Prototype-based Feature Representation Module (PFRM) to enhance robustness against visual noise and distortions under limited annotations. A dedicated 3D deviation estimation algorithm further enables accurate strabismus classification and angular measurement. Extensive experiments on the TongRenD and TEyeD datasets demonstrate that ProgNet outperforms state-of-the-art methods in segmentation accuracy. Clinical validation confirms that our system achieves high consistency with expert assessments, providing a standardized, non-invasive, and reliable solution for strabismus diagnosis.

源语言	英语
期刊	IEEE Journal of Biomedical and Health Informatics
DOI	https://doi.org/10.1109/JBHI.2026.3679532
出版状态	已接受/待刊 - 2026
已对外发布	是

访问文件

10.1109/JBHI.2026.3679532

其它文件与链接

链接到 Scopus 的出版物

引用此

@article{5bb67414e0d6446985ca00c5f67a2831,

title = "A VR-based Automated Strabismus Diagnosis System with Progressive Semi-Supervised Learning",

abstract = "Strabismus is a prevalent ocular disorder that can impair visual development and cause psychological issues if not diagnosed early. Conventional clinical diagnosis primarily relies on the prism cover test (PCT), which is subjective, requires patient cooperation, and lacks standardization. Recent advances in virtual reality (VR) and deep learning offer promising solutions for automated and standardized diagnosis. However, practical deployment faces three key challenges: realistic VR simulation of clinical exams, addressing image degradation (reflections/occlusions) with limited annotated data, and precise quantification of ocular deviations. In this study, we propose a novel VR-based automated strabismus diagnosis system by leveraging semi-supervised deep learning, and introduce a new clinical dataset, TongRenD. The framework incorporates five standardized clinical examination scenarios within a VR environment to ensure diagnostic consistency. We introduce ProgNet: an uncertainty-guided progressive semi-supervised segmentation network that integrates a Prototype-based Feature Representation Module (PFRM) to enhance robustness against visual noise and distortions under limited annotations. A dedicated 3D deviation estimation algorithm further enables accurate strabismus classification and angular measurement. Extensive experiments on the TongRenD and TEyeD datasets demonstrate that ProgNet outperforms state-of-the-art methods in segmentation accuracy. Clinical validation confirms that our system achieves high consistency with expert assessments, providing a standardized, non-invasive, and reliable solution for strabismus diagnosis.",

keywords = "Ocular image segmentation, Semi-supervised learning, Strabismus diagnosis, Virtual reality",

author = "Dehui Qiu and Bowei Ma and Ze Xiong and Yuhao Wang and Liguo Deng and Longfei Zhou and Xiaojie Cao and Weiwei Chen and Yumeng Wang and Meixia Jiang and Jing Fu and Fa Zhang and Xiaohua Wan",

note = "Publisher Copyright: {\textcopyright} 2013 IEEE.",

year = "2026",

doi = "10.1109/JBHI.2026.3679532",

language = "English",

journal = "IEEE Journal of Biomedical and Health Informatics",

issn = "2168-2194",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - JOUR

T1 - A VR-based Automated Strabismus Diagnosis System with Progressive Semi-Supervised Learning

AU - Qiu, Dehui

AU - Ma, Bowei

AU - Xiong, Ze

AU - Wang, Yuhao

AU - Deng, Liguo

AU - Zhou, Longfei

AU - Cao, Xiaojie

AU - Chen, Weiwei

AU - Wang, Yumeng

AU - Jiang, Meixia

AU - Fu, Jing

AU - Zhang, Fa

AU - Wan, Xiaohua

PY - 2026

Y1 - 2026

N2 - Strabismus is a prevalent ocular disorder that can impair visual development and cause psychological issues if not diagnosed early. Conventional clinical diagnosis primarily relies on the prism cover test (PCT), which is subjective, requires patient cooperation, and lacks standardization. Recent advances in virtual reality (VR) and deep learning offer promising solutions for automated and standardized diagnosis. However, practical deployment faces three key challenges: realistic VR simulation of clinical exams, addressing image degradation (reflections/occlusions) with limited annotated data, and precise quantification of ocular deviations. In this study, we propose a novel VR-based automated strabismus diagnosis system by leveraging semi-supervised deep learning, and introduce a new clinical dataset, TongRenD. The framework incorporates five standardized clinical examination scenarios within a VR environment to ensure diagnostic consistency. We introduce ProgNet: an uncertainty-guided progressive semi-supervised segmentation network that integrates a Prototype-based Feature Representation Module (PFRM) to enhance robustness against visual noise and distortions under limited annotations. A dedicated 3D deviation estimation algorithm further enables accurate strabismus classification and angular measurement. Extensive experiments on the TongRenD and TEyeD datasets demonstrate that ProgNet outperforms state-of-the-art methods in segmentation accuracy. Clinical validation confirms that our system achieves high consistency with expert assessments, providing a standardized, non-invasive, and reliable solution for strabismus diagnosis.

AB - Strabismus is a prevalent ocular disorder that can impair visual development and cause psychological issues if not diagnosed early. Conventional clinical diagnosis primarily relies on the prism cover test (PCT), which is subjective, requires patient cooperation, and lacks standardization. Recent advances in virtual reality (VR) and deep learning offer promising solutions for automated and standardized diagnosis. However, practical deployment faces three key challenges: realistic VR simulation of clinical exams, addressing image degradation (reflections/occlusions) with limited annotated data, and precise quantification of ocular deviations. In this study, we propose a novel VR-based automated strabismus diagnosis system by leveraging semi-supervised deep learning, and introduce a new clinical dataset, TongRenD. The framework incorporates five standardized clinical examination scenarios within a VR environment to ensure diagnostic consistency. We introduce ProgNet: an uncertainty-guided progressive semi-supervised segmentation network that integrates a Prototype-based Feature Representation Module (PFRM) to enhance robustness against visual noise and distortions under limited annotations. A dedicated 3D deviation estimation algorithm further enables accurate strabismus classification and angular measurement. Extensive experiments on the TongRenD and TEyeD datasets demonstrate that ProgNet outperforms state-of-the-art methods in segmentation accuracy. Clinical validation confirms that our system achieves high consistency with expert assessments, providing a standardized, non-invasive, and reliable solution for strabismus diagnosis.

KW - Ocular image segmentation

KW - Semi-supervised learning

KW - Strabismus diagnosis

KW - Virtual reality

UR - https://www.scopus.com/pages/publications/105034685008

U2 - 10.1109/JBHI.2026.3679532

DO - 10.1109/JBHI.2026.3679532

M3 - Article

AN - SCOPUS:105034685008

SN - 2168-2194

JO - IEEE Journal of Biomedical and Health Informatics

JF - IEEE Journal of Biomedical and Health Informatics

ER -

A VR-based Automated Strabismus Diagnosis System with Progressive Semi-Supervised Learning

摘要

访问文件

其它文件与链接

学术指纹

引用此