ROPFI

Analysis of Fundus Images of Retinopathy of Prematurity

What do we do? This project aims to aid computational support in the Retinopathy of the Prematurity (ROP) diagnosis through the analysis of Fundus Images. Some algorithms have been developed for it.

What is Fundus Image?

The fundus image is a picture of the retina. It is acquired by a fundus examination. This technique allows the inside of the eyeball to be observed to diagnose a disease or to check the evolution of pathologies.

What is ROP?

Retinopathy of the Prematurity (ROP) is a disease that develops in the retina of infants born prematurely or underweight. This pathology is defined by The International Classification of Retinopathy of Prematurity Revisited, DOI: 10.1001/archopht.123.7.991.

ROPFI Dataset

ROPFI has sixty-four images of ROP. Images were obtained during the period years 2013 and 2016 using the RetCam Shuttle camera (Clarity Medical Systems, Inc.). The dimensions of the images are 640 x 480 pixels, and their sizes are around 1 MB.

It is divided into four sub-datasets: images, manual_net, mask_1, and mask_2: images sub-dataset contains original fundus images of ROP, manual_net sub-dataset contains the vessels net labeled manually, mask_1 sub-dataset contains mask images to delimit retina region, and mask_2 sub-dataset contains mask images to delimit retina region where there are vessels.

Example of an image.


Sub-dataset images	Sub-dataset manual_net	Sub-dataset mask_1	Sub-dataset mask_2

You can use this dataset (ROPFI) citing the first publication bellow.

Publications:

1.-

2.- Monserrate Intriago-Pazmiño, Julio Ibarra-Fiallo, Raúl Alonso-Calvo, and José Crespo. 2019. Segmenting retinal vascular net from retinopathy of prematurity images using convolutional neural network. In Proceedings of the Second International Conference on Data Science, E-Learning and Information Systems (DATA ’19). Association for Computing Machinery, New York, NY, USA, Article 20, 1–5. DOI: 10.1145/3368691.3368711.

3.- Monserrate Intriago-Pazmiño, Julio Ibarra-Fiallo, Raúl Alonso-Calvo, José Crespo and Antonio Criollo-Ramos, "A new approach to two-dimensional filter for segmenting retinal vascular network from fundus images of premature born," 2017 IEEE International Symposium on Signal Processing and Information Technology (ISSPIT), Bilbao, 2017, pp. 405-409, DOI: 10.1109/ISSPIT.2017.8388677.

Areas:

Artificial Intelligence

Computer vision tasks

Applied computing

Computer informatics

Team

Head: Monserrate Intriago-Pazmiño, Ph.d. (c)

External researchers: Universidad Politécnica de Madrid, Madrid, Spain (Dr. José Crespo, Dr. Raúl Alonso-Calvo), Universidad San Francisco de Quito, Cumbayá, Ecuador (Prof. Julio Ibarra).

Pediatric ophthalmologists: Dr. Andrea Molinari and Dr. Monica Vargas, from Hospital Metropolitano. Dr. Jorge Luis Diaz and Dr. Fernando Diaz from Hospital Baca Ortiz. Both hospitals from Quito, Ecuador.

Computer Sciences Students: José Antonio Criollo, Michelle Sánchez.