Retinal Vasculature Segmentation Using Local Saliency Maps and Generative Adversarial Networks For Image Super Resolution
This work addresses the need for more accurate automated analysis of retinal images, particularly for small or blurred features, but is incremental as it builds on existing GAN and saliency techniques.
The authors tackled the problem of generating high-resolution retinal fundus images from low-resolution inputs using a GAN-based super-resolution method with local saliency maps, achieving perceptual quality close to original images and enabling retinal vasculature segmentation accuracy nearly matching that of original images at up to 16x scaling.
We propose an image super resolution(ISR) method using generative adversarial networks (GANs) that takes a low resolution input fundus image and generates a high resolution super resolved (SR) image upto scaling factor of $16$. This facilitates more accurate automated image analysis, especially for small or blurred landmarks and pathologies. Local saliency maps, which define each pixel's importance, are used to define a novel saliency loss in the GAN cost function. Experimental results show the resulting SR images have perceptual quality very close to the original images and perform better than competing methods that do not weigh pixels according to their importance. When used for retinal vasculature segmentation, our SR images result in accuracy levels close to those obtained when using the original images.