Hybrid fuzzy image processing for situation assessment [diabetic retinopathy]

Discusses a hybrid fuzzy image-processing system for situation assessment of diabetic retinopathy. The hybrid approach is motivated by the characteristics of the medical data and of the diagnostic decision-making process. The aim of the system is to support the early detection of diabetic retinopathy in a primary-care environment. For this purpose, both internal medicine. (diabetes) and ophthalmology have to be considered. The main input data are ophthalmological parameters, such as visual acuity, status of the anterior segment, status of the fundus, and previous therapies, and diabetological status; i.e., metabolic data. To reduce the huge number of parameters that have to be extracted by the ophthalmologist, image-processing methods for the automatic analysis of fundus photographs have been developed. The extraction is done by a multistage model-based approach. The segmentation results are used as an input to an overall fuzzy system that produces the final decision outcome (situation classes).