DiaReTDB1
The DiaReTDB1 dataset contains 89 images with an image size of 1500×1152 pixels. All the images were captured using the same 50-degree field-of-view digital fundus camera with varying imaging settings. 26 images contain exudates in the dataset with labled groundtruth.
来自论文”Detection of exudates in fundus photographs with imbalanced learning using conditional generative adversarial network”
HEI-MED
The HEI-MED dataset consists of 169 fundus images that are representative of a various degree of diabetic macular edema (DME) with a resolution of 2196×1958 pixels and with a 45 Field of View (FOV). All the images were captured with a Zeiss Visucam PRO fundus camera. 115 images are considered as normal retinal images and 54 retinal images are diagnosed with diabetic macular edema. 54 images contains exudates with labeled groundtruth while 115 images have no exudates.
来自论文”Detection of exudates in fundus photographs with imbalanced learning using conditional generative adversarial network”
MESSIDOR
which contains 1200 TIFF images with three different image sizes, 1440×960, 2240×1488 and 2304×1536 pixels. All the images were acquired using a color video 3CDD camera on a TopCon TRC NW6 with a 45-degree field of view. 800 images were acquired with pupil dilation and 400 without dilation. There are 226 images consisting of exudates and 974 images without exudates [26]. However, labeled exudates were not provided.
来自论文”Detection of exudates in fundus photographs with imbalanced learning using conditional generative adversarial network”
近期评论