Analogic preprocessing and segmentation algorithms for off-line handwriting recognition

Abstract

This report describes analogic algorithms used in the preprocessing and segmentation phase of offline handwriting recognition tasks. The handwriting recognition approach is segmentation based, i.e. it attempts to segment words into their constituent letters. In order to improve their speed the utilized CNN algorithms use dynamic, wave front propagation-based methods instead of relying on morphologic operators embedded into iterative algorithms. The system first locates handwritten lines in the page image then corrects their skew as necessary. Afterwards it searches for words within the lines and corrects skew at the word level as well. A novel trigger wave-based word segmentation algorithm is presented which operates on the skeletons of words. Sample results of experiments conducted on a database of 25 handwritten pages are presented.