The MNIST database of handwritten digits was provided by Yann LeCun when he was at Courant Institute, NYU, and by Corinna Cortes (Google Labs) and Christopher J.C. Burges (Microsoft Research). It is considered the standard for learning from real-world image data with minimal effort in preprocessing and formatting. The database consists of handwritten digits, offering a training set of 60,000 examples and a test set of 10,000. It is actually a subset of a larger set available from NIST. All the digits have been size-normalized and centered in a fixed-size image:
http://yann.lecun.com/exdb/mnist/
Figure 5: A sample of the original MNIST ...
