FLoRI21: Fluorescein Angiography Longitudinal Retinal Image Registration Dataset

Citation Author(s):
Li
Ding
University of Rochester
Tony
Kang
University of Rochester
Ajay
Kuriyan
Retina Service, Wills Eye Hospital
Rajeev
Ramchandran
University of Rochester
Charles
Wykoff
Houston Methodist Hospital
Gaurav
Sharma
University of Rochester
Submitted by:
Li Ding
Last updated:
Wed, 08/17/2022 - 10:38
DOI:
10.21227/ydp8-zf19
Data Format:
*.tif
License:
Creative Commons Attribution
1477 Views
Categories:
Standards Research Data
Image Processing
Medical Imaging
Computer Vision
Keywords:
fluoroscein angiography, image registration, retinal image analysis
0
0 ratings - Please login to submit your rating.
CITE SHARE/EMBED

Abstract 

The FLoRI21 dataset provides ultra-widefield fluorescein angiography (UWF FA) images for the development and evaluation of retinal image registration algorithms. Images are included across five subjects. For each subject, there is one montage FA image that serves as the common reference image for registration and a set of two or more individual ("raw") FA images (taken over multiple clinic visits) that are target images for registration.  Overall, these constitute 15 reference-target image pairs for image registration. Relevant details for the images can be found in the associated paper cited below. For each image pair, a set of manually selected corresponding points (control points) are included for quantitative evaluation of registration accuracy. A set of scripts is also included that illustrates the computation of accuracy metrics and visualizes the correspondences between control points in a reference and target image pair.

Instructions: 

Dataset

The naming convention of files are as follow:

Montage FA images: Montage_Subject_X.tif, where "X" is the subject ID.

Raw FA images: Raw_FA_N_Subject_X, where "N" is the raw FA image ID for this subject.

Ground truth control points: ControlPoints_Montage_FA_N_Subject_X.txt

 

The ground truth file for each pair of images has the following format:

Montage_point_1_x, Montage_point_1_y, FA_point_1_x, FA_point_1_y

Montage_point_2_x, Montage_point_2_y, FA_point_2_x, FA_point_2_y

...

Montage_point_M_x, Montage_point_M_y, FA_point_M_x, FA_point_M_y

 

Scripts

 Two demonstration scripts are included in this dataset.  The first one, "eval_registration_accuracy.m", calculates the registration success rate as a function of the mean registration error (MRE) threshold and the corresponding AUC number. The second one, "vis_control_points_FLoRI21.m",visualizes the correspondences between the control points in one raw FA image and the corresponding montage

 

Users of the dataset should cite the following paper

L. Ding et al., “Combining feature correspondence with parametric chamfer alignment: Hybrid two-stage registration for ultra-widefield retinal images,” IEEE Trans. Biomed. Eng., to appear. [Online]. Available: https://doi.org/10.1109/TBME.2022.3196458 

Dataset Files

LOGIN TO ACCESS DATASET FILES
Open Access dataset files are accessible to all logged in  users. Don't have a login?  Create a free IEEE account.  IEEE Membership is not required.

QUESTIONS?

Login to Send Author a Private Message

Report a problem with this Dataset