The Lucerne University of Applied Sciences and Arts (HSLU) is working with international partners to develop an AI-based app for the early detection of neglected tropical skin diseases (NTDs) in sub-Saharan Africa. To this end, images of dark skin types that are severely underrepresented in existing databases are being collected and used to train the AI. The aim of the project is to improve the diagnosis and monitoring of these diseases in regions where there is a lack of specialist personnel for skin diseases or where access to medical care is difficult.
Neglected tropical skin diseases (skin NTDs, see box) occur primarily in rural regions south of the Sahara (sub-Saharan Africa). They are often detected too late and can have serious health consequences and stigmatization if left untreated. This is where the "SkincAIr" project comes in: an international consortium, coordinated by the Universidad Politécnica de Madrid (Technical University of Madrid), is bringing together African and European partners to combat such diseases in a targeted manner. At the heart of the research project, which is funded by Horizon Europe, is an AI developed by Lucerne University of Applied Sciences and Arts (HSLU) that analyzes skin lesions on smartphone images and supports local healthcare professionals in making a diagnosis. Part of the project is also the targeted collection of image material for training the AI, as little data is currently available for darker skin types.
Digital diagnostics for all skin types in remote regions
The image material collected for AI training comes from different regions and covers as many different skin types, age groups and cultural backgrounds as possible. This ensures that the most diverse groups of people can be diagnosed equally correctly.Another advantage for remote regions: The app is designed as free software and can also be used without an internet connection. All collected data is automatically synchronized as soon as a connection is available again.
"Our AI will help to detect NTD skin diseases earlier, reduce the transmission rate and treat patients faster. This relieves the burden on the local healthcare infrastructure and significantly improves care," says Gil Sharvit, project manager at HSLU.
Successful test phase
The SkincAIr app is currently being tested in five countries: Kenya, Senegal, Ethiopia, Nigeria and the Democratic Republic of Congo. Local clinics and research institutions are supporting the collection of data for the app. So far, over 7000 images of skin diseases have already been carefully documented. The aim is to build up the largest open collection of skin images in sub-Saharan Africa and to ensure that the app works reliably in real clinical situations. Care is taken to ensure that the app is easy to use for local health professionals.Data is available to everyone
All data is shared in compliance with ethical standards, informed consent procedures and strict anonymization protocols. They are freely available to researchers worldwide with the aim of building the largest publicly accessible dataset on neglected tropical skin diseases. This will allow the solution to be further developed in other countries and extended to other diseases. In addition to AI, there is a focus on training local healthcare workers in order to improve care and strengthen the early detection of skin diseases in the long term.Neglected tropical diseases (NTD)
Neglected tropical diseases (NTDs) are a group of diseases that primarily affect poor people in tropical regions. In contrast to AIDS or malaria, NTDs receive less attention and research funding.Skin NTDs are a subgroup of these diseases that can manifest themselves through visible changes to the skin. These include parasite-borne diseases such as leishmaniasis, river blindness and scabies.
