This data set contains isolated handwritten characters of Gujarati. Isolated characters could be used in the development of many pattern recognition algorithms. Important considerations when evaluating the usefulness of a data set for experimental purposes include the number of samples it contains, the sampling rate at which it was captured, and the pixel depth at which the data is furnished. Considering all the mentioned parameters, 360 samples from different writers are collected for each character in Gujarati alphabet i.e. 34 consonants and 12 vowels. Thus this data set consists of around 12 K samples altogether. The characters are scanned at 300 dots per inch resolution. To enable the development of robust classifier characters have skew in them and noise pixels are included.