OCR dataset

The OCR dataset is a dataset of handwritten digits, each image is an 8x16 binary image, and there are 52152 samples in total.

BibTex: