Visual analysis for an early detection of diabetic neuropathy

Diabetic neuropathy is the most common long-term complication of diabetes mellitus. It comes along with significant nerve dysfunction, which is not reversible. Hence, it is essential to detect nerve fibre abnormalities as early as possible. In this paper, we investigate markers describing degradation of corneal nerves. We apply statistical computations and visual analysis to identify those variables of two clinical studies that separate DN patients from a control group. In this way, the diagnosis of DN patients is supported. The visual analysis is based on different representations visualizing both the statistical results and the gathered multi-variate data. The user can interactively manipulate the views, or select data that will be shown by further displays. In this way, the understanding of the data and its classification is supported. Ambiguous categorisations can be identified and grouped into a so-called "fuzzy group". For this group, further investigations are needed to decide about diabetic neuropathy.