With the help of related FAs, the OCT strips are created from OCT B-scans in four labels including MA, normal, abnormal, and vessel. Here, two datasets are prepared. In the first one, train, validation, and test sets are chosen randomly, while in the second one, the test set is isolated and test data is prepared from the images of cases that have not been included in the training and validation processes. In our first dataset, MA, normal, abnormal, and vessel classes comprise 87, 100, 72, and 131 strips of OCT images, respectively. The OCT strips have a width of approximately 31 pixels, whereas their height may vary. The images are cropped in a way that they contain only retinal layers while the other pixels are withdrawn. That is why the images have various heights.
Download Data:
Please reference the following paper if you would like to use any part of this dataset or method:
Almasi, R., Vafaei, A., Kazeminasab, E. et al. Automatic detection of microaneurysms in optical coherence tomography images of retina using convolutional neural networks and transfer learning. Sci Rep 12, 13975 (2022). https://doi.org/10.1038/s41598-022-18206-8