A Method for Pre-Compensation of Digital Images Based on Total Variation Deconvolution, and Comparing it with Wiener Deconvolution, to Enhance Visual Efficiency In The Presence Of Higher Order Ocular Optical Aberrations

Normal vision is essential to interact with widespread digital technologies. Higher order aberrations (HOA) are one of the important causes of image degradation in patients suffering from keratoconus, pterygium, or irregular astigmatism. These aberrations cannot be corrected by common ways of refractive errors correction like spectacles. Besides, other methods of HOA correction (e.g. customized contact lenses) are not easily accessible in many regions and photo-refractive surgeries like laser-assisted in situ keratomileusis (LASIK) are expensive. A novel method to correct HOA is pre-compensation of images displayed on computer monitors to provide unaberrated images. We used total variation (TV) deconvolution to pre-compensate images and compared the results with previous methods based on Wiener deconvolution.

MATLAB was used for simulations. Each HOA was reconstructed using single-index scheme for Zernike coefficients. Pupil diameter was assumed to be four millimeters. Root-mean-square error of simulated aberrations was greater than 0.35 micrometers. Point spread functions corresponding to each HOA were created and TV deconvolution was used to pre-compensate the images.

Strehl ratio of TV pre-compensation is very close to the results of Wiener pre-compensation method. Besides, execution time for TV Pre-compensation algorithm is about 20% faster than Wiener.

Simulations ascertain that TV based pre-compensation method is a fast and efficient way to pre-compensate the images, and it could improve the patient’s perceived retinal image quality. One more advantage of using TV algorithm is that we could enhance the results just by altering one parameter.