The chapter describes automated quantitative vascular biomarkers in retinal fundus images for the early warning of retinal damage, as developed in the Sino-Dutch RetinaCheck project. The software implements brain-inspired algorithms, exploiting multiscale, and multiorientation geometric methods. A full pipeline is presented, including background normalization, crossing-preserving vessel enhancement, nonlinear denoising, and contextual vessel completion. The quantitative biomarkers are vessel width, vessel tortuosity, vessel bifurcation measures, and several fractal dimensions. The methods are extensively validated and show high robustness and state-of-the-art performance.
|Title of host publication||Computational Retinal Image Analysis|
|Subtitle of host publication||Tools, Applications and Perspectives|
|Editors||Emanuele Trucco, Tom MacGillivray, Yanwu Xu|
|Publisher||Academic Press Inc.|
|Number of pages||34|
|Publication status||Published - 22 Nov 2019|
Huang, F., Abbasi-Sureshjani, S., Zhang, J., Bekkers, E. J., Dasht Bozorg, B., & ter Haar Romeny, B. M. (2019). Vascular biomarkers for diabetes and diabetic retinopathy screening. In E. Trucco, T. MacGillivray, & Y. Xu (Eds.), Computational Retinal Image Analysis: Tools, Applications and Perspectives (pp. 319-352). Academic Press Inc.. https://doi.org/10.1016/B978-0-08-102816-2.00017-4