An Off Line Cursive Script Recognition System Using Fourier -Wavelet Features

This paper describes a complete system for recognition of offline cursive handwriting. Preprocessing techniques, which include slant, slope, stroke thickness, segmentation, and normalisation of images are described. A new efficient feature extraction method based on Fourier wavelet transform is implemented and analyzed. The recognizer starts with features extracted in a coarse resolution and with successive passes renders the same features at finer resolution till the classification meets the acceptance criteria. This method is tested on a database of cursive script and is proven as an efficient representation compared to other features. Since Fourier wavelet descriptor is rotational invariant, this algorithm works for any style of handwriting.