Annotation-free 3D reconstruction and quantification of retinal microvasculature by RADAR - Takeaways - MDSpire

Annotation-free 3D reconstruction and quantification of retinal microvasculature by RADAR

  • By

  • Hao Zhang

  • Xindi Liu

  • Jiayi Wu

  • Ke Wu

  • Dong Li

  • Bin Chen

  • Hong Wang

  • Jinyang Liang

  • Zhanle Lin

  • Yuping Zheng

  • Liang Yao

  • January 21, 2026

  • 0 min

Share

  • 1

    RADAR is an annotation-free framework for 3D segmentation and quantification of retinal microvasculature using optical coherence tomography angiography.

  • 2

    The framework integrates adaptive physics-aware denoising and topology-preserving centerline extraction for accurate network reconstruction.

  • 3

    RADAR was validated in healthy individuals and early-stage diabetic retinopathy patients, outperforming standard segmentation tools.

  • 4

    Quantitative analysis with RADAR revealed compensatory remodeling and increased tortuosity in diabetic eyes, highlighting its diagnostic potential.

  • 5

    RADAR enables precise extraction of volumetric biomarkers, facilitating early detection and monitoring of ocular and systemic vascular diseases.

Original Source(s)

Annotation-free 3D reconstruction and quantification of retinal microvasculature by RADAR

  • by Hao Zhang, Xindi Liu, Jiayi Wu, Ke Wu, Dong Li, Bin Chen, Hong Wang, Jinyang Liang, Zhanle Lin, Yuping Zheng, Liang Yao

  • January 21, 2026

Related Content

Evaluating Retinal Vascular Oxygen Saturation Levels in Myopic Children Versus Adults

Floaters May Signal Higher Risk Than Flashes

New study examines retinal detachment risk in primary care presentations

Inner retinal layer thickness reflects plasma biomarkers in preclinical Alzheimer’s disease