Distributed Skin Lesion Analysis Across Decentralised Data Sources

Skin cancer has become the most common type. Research applied image processing and analysis tools to support improve diagnose process. Conventional procedures usually centralise data from various sources a single location execute tasks on central servers. However, centralisation of medical does not often comply with local protection regulations due its sensitive nature loss sovereignty if providers allow unlimited access data. The Personal Health Train (PHT) is Distributed Analytics (DA) infrastructure bringing algorithms instead vice versa. By following this paradigm shift, it proposes solution for persistent privacy- related challenges. In work, we present feasibility study, which demonstrates capability PHT perform statistical analyses Machine Learning skin lesion distributed among three Germany-wide providers.