New AI Dataset Advances Dermatology for Darker Skin Tones

In sub-Saharan Africa, where dermatological resources are critically scarce, untreated skin conditions affect millions of people, particularly children. A new initiative led by researchers at the University of Basel, in collaboration with dermatologists in Madagascar, Malawi, and Guinea, seeks to address this disparity. By creating a robust image dataset specifically tailored to darker skin tones, the PASSION project is laying the groundwork for artificial intelligence to play a significant role in improving dermatological diagnostics in underserved regions.

Filling the Data Gap for Darker Skin Tones

One major obstacle to AI-driven dermatology is the lack of image datasets representing skin conditions in darker skin tones. Current AI models rely heavily on photos of skin diseases in lighter-skinned patients, primarily sourced from Europe and the U.S. This lack of diversity in training data limits the diagnostic accuracy of AI tools for patients with pigmented skin, leaving an already underserved population further marginalized.

The PASSION project (Pediatric AI Skin Support In Outreach Nations) addresses this gap by compiling a database of over 4,200 annotated images depicting common skin diseases in darker skin tones, such as eczema, fungal infections, scabies, and bacterial skin infections. These images, collected with patient consent between 2020 and 2023, include data from more than 1,300 patients, two-thirds of whom are under 18. The dataset is designed not only to train new AI models but also to evaluate the accuracy of existing systems, ensuring more inclusive diagnostic tools.

Why It Matters: Addressing a Dermatology Crisis

The urgency for solutions is clear. In many African nations, there are fewer than one dermatologist per million people—far below the World Health Organization’s recommendation of one per 50,000. This shortage is particularly acute in rural regions, where up to 87% of children suffer from untreated skin diseases. Without intervention, these conditions can lead to significant morbidity, affecting overall quality of life.

By leveraging AI to assist with diagnostics, the PASSION project envisions a future where patients can use smartphones to photograph their skin conditions and receive treatment recommendations directly from AI. As Professor Alexander Navarini explains, “Our vision is that each patient will be able to take a photo of their skin problem themselves free of charge using a smartphone and then upload it. They will then receive a treatment recommendation from AI.” Dermatologists would step in only if the problem persists or fails to improve.

Early testing is already underway as part of a validation study in Madagascar. According to Philippe Gottfrois, a doctoral student and lead author of the study, “We are currently testing the method step by step as part of a validation study in Madagascar. Once diagnostic accuracy exceeds 80%, we intend to offer the new diagnostic tool with scientific monitoring.”

Expanding the Dataset to Bridge Care Gaps

Looking ahead, researchers plan to broaden the dataset to include neglected tropical skin diseases, which are highly prevalent in these regions but remain underrepresented in diagnostic tools. By addressing this gap, the team hopes to narrow the stark inequities in dermatological care and bring innovative, scalable solutions to underserved populations.

With its focus on inclusivity and accessibility, the PASSION project exemplifies how technology can advance global health and address care disparities in resource-limited settings. As AI-based tools continue to evolve, they hold promise as a transformative force for improving healthcare outcomes worldwide.

To learn even more about the use of AI in dermatology, check out the "How Working with AI May Make Us Better Dermatologists" session at the 2025 Dermatology Foundation Clinical Symposium.