Friedrich-Alexander-Universität Erlangen
Lehrstuhl für Mustererkennung
Martensstraße 3
91058 Erlangen
The public database contains at the moment 15 images of healthy patients, 15 images of patients with diabetic retinopathy and 15 images of glaucomatous patients. Binary gold standard vessel segmentation images are available for each image. Also the masks determining field of view (FOV) are provided for particular datasets. The gold standard data is generated by a group of experts working in the field of retinal image analysis and clinicians from the cooperated ophthalmology clinics. We intend to add further gold standard data to the existing images to help the evaluation of algorithms which localize the macula, optic disc, or differentiate between arteries and veins.
Download the whole dataset (~73 Mb)
Input Data:
Download input images of healthy eyes (~18 Mb)Download input images of glaucomatous eyes (~21 Mb)Download input images of eyes with diabetic retinopathy (~28 Mb)
Field of View (FOV) masks:Download the FOV mask images for the healthy eyes (~1 Mb)Download the FOV mask images for the glaucomatous eyes (~1 Mb) Download the FOV mask images for the eyes with diabetic retinopathy (~1 Mb)
Vessel Segmentation Goldstandard:
Download manually labeled images of healthy eyes (~2 Mb)Download manually labeled images of glaucomatous eyes (~2 Mb)Download manually labeled images of eyes with diabetic retinopathy (~2 Mb)
Optic Disk Goldstandard (Center Point and Radius):
Download gold standard data of two experts for optic disk localization
Preliminary Version:
We captured 18 image pairs of the same eye from 18 human subjects using a Canon CR-1 fundus camera with a field of view of 45° and different acquisition setting. For each pair, the first image has poor quality and thus the examination had to be repeated. Both images share approximately the same field of view, whereas small shifts were caused by eye movements between the acquisitions.
In the current version of our database, images of poor quality suffer from decreased sharpness (locally or globally), e.g. due to a defocused camera. Further quality features (e.g. image contrast or illumination conditions) may be considered in a future version. The available dataset was captured by Jan Odstrcilik. If you use it in your publications, please cite:
Thomas Köhler, Attila Budai, Martin Kraus, Jan Odstrcilik, Georg Michelson, Joachim Hornegger. Automatic No-Reference Quality Assessment for Retinal Fundus Images Using Vessel Segmentation, 26th IEEE Internatioal Symposium on Computer-Based Medical Systems 2013, Porto
(to appear)
Download whole datasetDownload images of good qualityDownload images of poor quality
This database has been established by a collaborative research group to support comparative studies on automatic segmentation algorithms on retinal fundus images. The database will be iteratively extended and the webpage will be improved.
We would like to help researchers in the evaluation of segmentation algorithms. We encourage anyone working with segmentation algorithms who found our database useful to send us their evaluation results with a reference to a paper where it is described. This way we can extend our database of algorithms with the given results to keep it always up-to-date.
The database can be used freely for research purposes. We release it under Creative Commons 4.0 Attribution License. If you are using our database to evaluate your methods, please cite
Budai, Attila; Bock, Rüdiger; Maier, Andreas; Hornegger, Joachim; Michelson, Georg. Robust Vessel Segmentation in Fundus Images. International Journal of Biomedical Imaging, vol. 2013, 2013
The public database contains at the moment 15 images of healthy patients, 15 images of patients with diabetic retinopathy and 15 images of glaucomatous patients. Binary gold standard vessel segmentation images are available for each image. Also the masks determining field of view (FOV) are provided for particular datasets. The gold standard data is generated by a group of experts working in the field of retinal image analysis and clinicians from the cooperated ophthalmology clinics. We intend to add further gold standard data to the existing images to help the evaluation of algorithms which localize the macula, optic disc, or differentiate between arteries and veins.
Download the whole dataset (~73 Mb)
Input Data:
Download input images of healthy eyes (~18 Mb)Download input images of glaucomatous eyes (~21 Mb)Download input images of eyes with diabetic retinopathy (~28 Mb)
Field of View (FOV) masks:Download the FOV mask images for the healthy eyes (~1 Mb)Download the FOV mask images for the glaucomatous eyes (~1 Mb) Download the FOV mask images for the eyes with diabetic retinopathy (~1 Mb)
Vessel Segmentation Goldstandard:
Download manually labeled images of healthy eyes (~2 Mb)Download manually labeled images of glaucomatous eyes (~2 Mb)Download manually labeled images of eyes with diabetic retinopathy (~2 Mb)
Optic Disk Goldstandard (Center Point and Radius):
Download gold standard data of two experts for optic disk localization
Preliminary Version:
We captured 18 image pairs of the same eye from 18 human subjects using a Canon CR-1 fundus camera with a field of view of 45° and different acquisition setting. For each pair, the first image has poor quality and thus the examination had to be repeated. Both images share approximately the same field of view, whereas small shifts were caused by eye movements between the acquisitions.
In the current version of our database, images of poor quality suffer from decreased sharpness (locally or globally), e.g. due to a defocused camera. Further quality features (e.g. image contrast or illumination conditions) may be considered in a future version. The available dataset was captured by Jan Odstrcilik. If you use it in your publications, please cite:
Thomas Köhler, Attila Budai, Martin Kraus, Jan Odstrcilik, Georg Michelson, Joachim Hornegger. Automatic No-Reference Quality Assessment for Retinal Fundus Images Using Vessel Segmentation, 26th IEEE Internatioal Symposium on Computer-Based Medical Systems 2013, Porto
(to appear)
Download whole datasetDownload images of good qualityDownload images of poor quality
A test version of an applet is available here to speed up the uploading and evaluation of images
[odstrcilik09] |
||||||
[odstrcilik09] |
||||||
[odstrcilik09] |
[budai09] |
|||||
[odstrcilik09] |
|||||
(Everything is background) |
Responsible person for maintaining this homepage is:
Responsible persons for the database are:
The database is provided by the Pattern Recognition Lab (CS5), the Department of Ophthalmology, Friedrich-Alexander University Erlangen-Nuremberg (Germany), and the Brno University of Technology, Faculty of Electrical Engineering and Comunnication, Department of Biomedical Engineering, Brno (Czech Republic).
This work has been supported by the national research center DAR (Data, Algorithms and Decision making) project no. 1M0572 coordinated by the Institute of Information Theory and Automation, Academy of Science, Czech Rep. and partly also by the institutional research frame no. MSM 0021630513; both grants sponsored by the Ministry of Education of the Czech Republic. The authors highly acknowledge the cooperation with the Eye Clinic Zlin, Czech Rep. (T. Kubena, M.D. and P. Cernosek, MSc), through which also the test set of images was provided.
Attila Budai is supported by the International Max Planck Research School for Optics and Imaging.
The authors gratefully acknowledge funding of the Erlangen Graduate School in Advanced Optical Technologies (SAOT) by the German National Science Foundation (DFG) in the framework of the excellence initiative.
odstrcilik09:
Jan Odstrcilik, Jiri Jan, Radim Kolar, and Jiri Gazarek. Improvement of vessel segmentation by matched filtering in colour retinal images. In IFMBE Proceedings of World Congress on Medical Physics and Biomedical Engineering, pages 327 - 330, 2009.
DOI: 10.1049/iet-ipr.2012.0455, Print ISSN 1751-9659, Online ISSN 1751-9667
Earlier stages of the datasets are available here