KMNIST

The dataset used for the Drawing out of Distribution (DooD) model, which consists of stroke-based images of handwritten Japanese characters.

BibTex: