Automated anterior chamber angle pigmentation analyses using 360° gonioscopy.

PURPOSE: To assess the pigmentation distribution in the iridocorneal angle using an established algorithm with gonioscopically obtained images. METHODS: Manual and automatically modified Scheie's pigmentation grading (ie, 0/I=0, II=1 and III/IV=2) of trabecular meshwork was performed using an established algorithm on the 75 open-angle eyes of 75 subjects obtained by automated gonioscopy. All images were collected at the Matsue Red Cross Hospital in 2016. The differences in the pigmentation density were compared statistically between the automated and manual techniques and among the four sectors (ie, inferior, superior, temporal and nasal) and the four quadrants. RESULTS: There was substantial agreement between both grading methods (kappa value=0.70). There was no significant difference between the automated and manual grading in any sectors except for the superior (p=0.0004). The automated pigmentation grade was significantly (p<0.05) higher in the inferior sector (mean grade, 1.43) than in the others (mean grade, 0.48~0.76), and it was also significantly (p<0.05) higher in the inferior quadrant (mean grade, 3.56) than in the others (mean grade, 1.64~2.24). The findings were similar for manual grading. CONCLUSIONS: The entire distribution of the pigmentation in the anterior chamber angle was classified successfully using the algorithm, and the automated versus manual grading comparison showed good agreement. The automated pigmentation grading scores in the inferior sector and inferior quadrant were significantly higher than in the others as previously reported. Similar findings also were seen for manual grading.

Model based retinal analysis for retinopathy detection.

This paper provides a novel combination of methods to preprocess and extract features from retinal images, in order to get structures such as the vessel tree and the optic disc, and diagnose retinopathy. A model based approach was developed, making extensive use of LBG vector quantization and elements provided by mathematical morphology as investigation tools. The proposed algorithm was tested against the DRIVE database hand-labeled ground-truth, scoring an overall 93.2% correspondence on vessel tree search algorithm, 81.3% on the optic disc localization, 85% specificity and 78% correspondence on the disease detection.