Premium
Custom‐devised and generic digital enhancement of images for people with maculopathy
Author(s) -
Leat Susan J.,
Mei Ming
Publication year - 2009
Publication title -
ophthalmic and physiological optics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.147
H-Index - 66
eISSN - 1475-1313
pISSN - 0275-5408
DOI - 10.1111/j.1475-1313.2008.00633.x
Subject(s) - contrast (vision) , maculopathy , artificial intelligence , macular degeneration , computer vision , filter (signal processing) , sharpening , age related maculopathy , visibility , computer science , mathematics , ophthalmology , optics , medicine , physics , retinopathy , diabetes mellitus , endocrinology
Aim: The purpose of this study was to compare the effectivity, in terms of the potential usefulness, of digital filters based on either contrast sensitivity (CS) or supra‐threshold contrast matching (CM) in enhancing pictures images for people with maculopathy and to investigate whether generic filters (not based on an individual’s vision loss) are equally as effective. Effectivity is measured by changes in perceived visibility. Methods: Thirty‐five subjects with maculopathy, aged 20–92 years, took part [13 atrophic age‐related macular degeneration (ARMD), 14 exudative ARMD, and 8 juvenile macular dystrophy (JMD)]. CS and supra‐threshold CM were measured. A range of CS filters (1 or 2‐octave wide band‐pass filter using a Gabor or polynomial envelope) with different strengths were developed based on the ratio of the individual’s contrast threshold and that of a normal age‐related group. Similarly filters were developed based on CM at 3.6% and 27.9% contrast. The following generic filters were also applied with different ‘strengths’: edge enhancement; sharpening; contrast enhancement; Peli’s adaptive enhancement; difference of Gaussian; and an equi‐emphasis band‐pass filter. The filters were applied to images of faces and general scenes. Subjects were asked to rank the perceived visibility of images (to obtain the best version of each filter) and then to rate the perceived visibility of each image filtered with a particular filter. Results: In general, subjects with atrophic ARMD and JMD preferred the weaker versions of most of the filters, while those with exudative ARMD did not show such a clear preference. Generally, images of faces were preferred with less enhancement than scenes. The filters based on CM were rated as giving significant improvement, while those based on CS and peak emphasis were not preferred. Of the generic filters, the Peli adaptive enhancement filter was most frequently rated as giving a significant improvement ( p < 0.05) followed by the contrast enhancement filter. They gave the same perceived enhancement as the custom‐devised filters. Conclusions: Generic filters, which are easier to apply than the custom‐devised filters, are appropriate for rehabilitation purposes.