Handwritten script recognition using DCT, Gabor filter, and wavelet features at word level

Abstract

In a country like India, many of the documents such as office letters, checks, envelopes, forms, and other types of manuscripts are multiscript in nature. A document consisting of English script and a regional script is quite common. Hence, automatic recognition of scripts present in a multiscript document has a variety of practical and commercial applications in banks, post offices, reservation counters, libraries, etc. In this paper, a multiple feature-based approach is presented to identify the script type from a multiscript document. Features are extracted using Gabor filters, discrete cosine Transform, and wavelets of Daubechies family. Nine popular Indian scripts are considered for recognition in this paper. Experiments are performed to test the recognition accuracy of the proposed system at word level for bilingual scripts. Using neural network classifier, the average success rate is found to be 97 %. © 2013 Springer.