Fundus image quality assessment for an eye model with cataract using adapted illumination patterns

Purpose In recent years, different methods have been developed to improve fundus image quality in eyes with cataract. For stray light reduction, we have introduced a segmented LED light source. To analyze the effect of different illumination patterns at different cataract characteristics, we developed an optical cataract eye model. Accordingly, we performed fundus imaging using a fundus camera equipped with a LED light source controlled by a Raspberry Pi. Methods A customized 24‐fold segmented LED light source was connected to a fundus camera. The eye model consists of a plano‐convex lens, a cover glass, and a test chart with five black‐white test fields. The pupil diameter was set to 7 mm. To test the influence of variable illumination patterns, four different cortical cataract patterns were produced by sand blasting the cover glasses. Three spoke‐like (1Spoke, 4S, 7S) patterns and one crescent‐shaped/CS pattern were realized. To analyze the effect of adapted illumination conditions, the gray‐level co‐occurrence matrix (GLCM) contrasts were calculated. At that, one image was taken at standard annular illumination (all LEDs on) as baseline followed by seven defined adapted configurations, for each type of cataract. Results For the four types of cataracts, the baseline GLCM contrast values were 0.21 (1S), 0.18 (4S), 0.17 (7S), and 0.22 (CS). In each cataract type, the best‐fitting illumination pattern resulted in 0.23 (1S), 0.29 (4S), 0.22 (7S), and 0.23 (CS). Compared to the baseline values, the increases of the GLCM contrasts were 9.5% (1S), 61% (4S), 29% (7S), and 4.5% (CS). Conclusions We successfully performed structured LED illumination for different cortical cataract types using an optical eye model. An improvement of fundus image quality could be determined in all investigated cases using the gray‐level co‐occurrence matrix contrast.