Artificial Intelligence (AI) has shown promise in dermatology , but most models are trained on Western datasets—leaving Indian skin tones poorly represented. The newly developed DermaCon-IN dataset (5,450+ images, 240+ diagnoses) could change this by capturing real-world presentations from Indian OPDs. Why does this matter? Conditions like melasma, vitiligo, and fungal infections often look different on darker skin, leading to misclassification by Western-trained AI.
By reflecting our spectrum of tones and disease patterns, DermaCon -IN enables algorithms that “see what we see.” For dermatologists, this could mean faster triage, fewer diagnostic misses, and research tools tuned for Indian practice. Still, concerns around over-reliance on AI, cost, and privacy remain. But for the first time, AI in dermatology feels closer to our clinics than our textbooks. Would you trust an AI that…
