ROSE-O: A retinal structure detection dataset of OCTA images
This dataset is released for academic research use only.
Dataset download:
https://doi.org/10.5281/zenodo.12775880
Codes download:
The code of our Voting-based Adaptive Feature Fusion multi-task network (VAFF-Net) can be found here:
https://github.com/iMED-Lab/VAFF-Net
Related papers：https://ieeexplore.ieee.org/document/9870738
[1] Jinkui Hao et al., "Retinal Structure Detection in OCTA Image via Voting-based Multi-task Learning," in IEEE Transactions on Medical Imaging, 2022, doi: 10.1109/TMI.2022.3202183.
Datatset Description
ROSE-O is a retinal structure detection dataset of OCTA images, with precise manual annotations of RV, RVJ and the FAZ. It contains 117 images which were captured using the Optovue Avanti RTVue XR with AngioVue software (Optovue, Fremont, USA): the images have a resolution of 304×304 pixels. The SVC, DVC and IVC angiograms of each participant were obtained by the device.
Examples of manual annotations from ROSE-O subset. Illustrations of the OCTA en face angiograms and typical retina structure. (A) 3D OCTA volume. (B) Retinal layer segmentation of inner and outer retina. (C) The en face angiograms of the inner vascular complexes (IVC),superficial vascular complex (SVC), and deep vascular complex (DVC), respectively. (D) Typical retina structures: RV, RVJ and FAZ.