Vitiligo is a common, acquired, chronic skin disease, which is characterized by hypopigmented (pale) areas of skin. If exposed areas such as the face is affected, it can be disfiguring. The areas often progress without therapy such as UV-B phototherapy or topical corticosteroids. This is why early and proper therapy is essential. To grade the severity of vitiligo, several scoring systems have been proposed, where the simplest system is the counting amount of skin in % which is affected (body surface area, BSA). Even though most scoring systems display good interrater variability, they are seldom used in clinical practice due to time constraints. This underlines the need for a more effective way to evaluate disease course and therapy efficacy in vitiligo. A 3D total body scanner captures almost the whole skin surface with visible light cameras in a quick and highly standardized manner. This is why, when paired with machine learning algorithms, its data offers the perfect base to automatize the scoring in vitiligo, and thus improve patient outcomes of this skin disease. Swarm learning refers to the training of a machine learning algorithm from multiple training sites without sharing the underlying data, and thus conforming to the strict data protection requirements of patient data. In this project, we want to establish swarm learning of an algorithm to score vitiligo from 3D body scanner data generated at two university hospitals in Germany.