The MNIST database (Modified National Institute of Standards and Technology database) is a large database of handwritten digits that is very popularly used for training image processing systems and computer vision techniques. Since its release in 1999, this classic dataset of handwritten images has served as the basis for benchmarking classification algorithms.
The database contains 70000 digitally scanned images (60000 training and 10000 testing) of handwritted digits (0-9). Digit recognition is done using the 784 pre-extracted features that correspond to pixel intensities of a 28x28 grid.
Details on the dataset, algorithms that have been attempted and their success levels can be found on http://yann.lecun.com/exdb/mnist/index.html.
Kaggle page for the ongoing MNIST competetion: https://www.kaggle.com/c/digit-recognizer
Wikipedia page on the MNIST database: https://en.wikipedia.org/wiki/MNIST_database