Image Correction for Improving Visual Acuity Using Zernike-Based Vision Simulation

This paper proposes a novel image correction method to improve visions of individuals with refractive errors such as myopia, hyperopia, and astigmatism. Refractive errors can be corrected by eyeglasses, contact lenses, or surgery, but there are problems such as maintenance or cost. Our method corrects images so that the corrected images are perceived similarly to the original images without eyeglasses. Our method consists of two models: image correction model and blur simulation model. The image correction model is a convolutional neural network model designed for super-resolution tasks, and the blur simulation model simulates the perceived blur of a human eye. Perceived blur can be modeled by Zernike polynomials-based point spread function (PSF), and blurred images are computed by convolving the images with PSF. Our model works as follows. First, images are fed to the image correction model to produce corrected images. Then, these corrected images are passed to the blur simulation model. The image correction model is trained so that blurred corrected images become identical to the original images. After training is completed, only the first step is required to correct images. Our approach can correct any type of image using the same model without significant contrast loss.