Deep learning can generate traditional retinal fundus photographs using ultra-widefield images via generative adversarial networks.
Tae Keun YooIk Hee RyuJin Kuk KimIn Sik LeeJung Sub KimHong Kyu KimJoon Yul ChoiPublished in: Comput. Methods Programs Biomed. (2020)
Keyphrases
- retinal fundus
- deep learning
- three dimensional
- image features
- object recognition
- image classification
- phase congruency
- image registration
- input image
- generative model
- image retrieval
- segmentation algorithm
- planar surfaces
- multiple images
- image processing
- keypoints
- retinal images
- image structure
- infrared images
- single image
- unsupervised learning