Synthetic HAM10000 Extension Dataset with Skin Tone Balance

This dataset presents a synthetic collection of skin lesion images generated using a tone-conditioned Generative Adversarial Network (GAN) to address bias and underrepresentation of darker skin tones in dermatological datasets. The dataset includes 10,000 images equally divided between medium and dark skin tones, reflecting the lesion class distribution of the original HAM10000 dataset. This work contributes toward AI fairness by providing balanced and realistic lesion samples for machine learning model development, augmentation, and evaluation.