.webp&w=384&q=75){kind=logo}
Loading, please wait...
Loading, please wait...
Researchers recently introduced CUNEX. Specifically, this is an innovative open-source corneal segmentation deep learning framework designed for anterior segment optical coherence tomography (AS-OCT). In addition, this tool automates the segmentation process. Consequently, it is essential for diagnosing corneal pathologies. Furthermore, by utilizing a large dataset from Moorfields, the model identifies healthy corneas. Moreover, it detects conditions like keratoconus. Therefore, this advancement addresses the need for reliable analysis. In conclusion, it represents a major step forward.
Furthermore, CUNEX outperforms traditional models. Specifically, it generalizes across different devices. However, most existing algorithms suffer from performance drops when using new hardware. In contrast, CUNEX maintains high accuracy across multiple platforms. Similarly, it works well with Casia and Anterion devices. Therefore, it provides a stable foundation for research. Moreover, the open-source nature encourages transparency. Consequently, it fosters rapid improvement. In addition, it supports multicenter research. Finally, it ensures data consistency.
Additionally, the accuracy of this framework allows for precise tasks. For instance, it improves disease staging for Fuchs endothelial corneal dystrophy. Moreover, the model helps analyze demographic factors like age and sex. Because the segmentation is meaningful, it ensures clinical relevance. Specifically, clinicians can rely on automated measurements. Consequently, they can monitor progression. Therefore, this efficiency improves management. In addition, it boosts diagnostic accuracy. Furthermore, it saves time for practitioners. As a result, patient outcomes may improve.
CUNEX offers superior cross-device external validation. Specifically, while other models degrade when using images from different OCT machines, CUNEX maintains high accuracy. Therefore, it is more versatile for various clinical settings.
The model was specifically trained on healthy eyes, keratoconus, and Fuchs endothelial corneal dystrophy. Consequently, it provides robust segmentation for these common conditions.
Disclaimer: This content is for informational and educational purposes only. It does not constitute medical advice or a professional relationship. Refer to the latest local and national guidelines for clinical practice.
References
Kandakji L et al. An Open-Source Deep Learning Framework for Automated Corneal Segmentation in Anterior Segment Optical Coherence Tomography With Cross-Device External Validation. Cornea. 2026 Apr 10. doi: 10.1097/ICO.0000000000004130. PMID: 41962147.
Falk T, et al. U-Net: deep learning for cell counting, detection, and morphometry. Nature Methods. 2019.
Read summarized clinical updates, watch expert medical content, and earn CME certifications right from your smartphone.
Open your smartphone camera to scan and install.
CUNEX is an open-source deep learning framework providing automated corneal segmentation in AS-OCT images with high accuracy across different devices....
3 months ago
Andhra Pradesh reported 10 new Covid-19 cases, taking the state tally to 49 while deaths remain at four. With 24 patients hospitalized and 16 under home isolation, the Health Department has intensified monitoring. Medical professionals should review regional distribution, diagnostic protocols, and management plans.
Today
An 11-year Swedish registry study of 618 uterine sarcoma patients found that minimally invasive surgery yielded survival comparable to open surgery in early stages. However, adjuvant chemotherapy conferred no survival benefit in localized or advanced disease, highlighting stage and histology as key outcomes.
3 days back
A cross-sectional study evaluates post-intensive care syndrome in cardiac patients 2-4 weeks post-ICU discharge, highlighting cognitive, psychological, and functional impairments and the need for structured multidisciplinary rehabilitation.
3 days back
Anterior cruciate ligament reconstruction failure lacks uniform definition. A narrative review proposes an integrative framework incorporating objective and subjective instability, persistent pain, restricted motion, graft rupture, and secondary meniscal injury to standardize clinical reporting.
3 days back
With World Obesity Atlas data warning that over 41 million Indian children are overweight or obese, ICMR and NIN have unveiled a 10-point policy roadmap. The initiative calls for mandatory front-of-pack labeling, HFSS taxes, strict marketing bans, and healthier school environments to curb non-communicable diseases.
Today
