CHoiCe: A Complex Handwritten Character dataset

The data of the dataset is collected from Professor Tom Gedeon and the complete handwriting paper of the CEDAR handwriting dataset. A CHoiCE Dataset with 62 classes cursive handwriting letters, "0-9, a-z, A-Z",
each class in both the original data and the binary data at least have 40 pictures. The data format is a 28x28 ".png" format picture. The data set has a total of 62 categories of 0-9, a-z and A-Z, corresponding to the files "0" to "61" in the order of "label.txt". The data set is divided into two parts, the unprocessed original data image is stored in the "0" to "61" in the "V0.3/data" folder, and the binarized data image Stored in "0" to "61" in the "V0.3/data-bin" folder.