Feature extraction based on fuzzy set theory for handwriting recognition

This paper presents a method based on fuzzy set theory for extracting features from handwriting words. After the feature extraction and the word segmentation process a handwriting word is represented by an ordered sequence of line segments. For each of these segments are calculated memberships values to fuzzy sets representing different types of curve lines and straight lines. The position of the line segments in a letter or piece of letter resulting from the word segmentation is also evaluated by means of fuzzy sets. Frizzy Hidden Markov Models are employed to classify the handwriting words. A database compounded by handwriting words extracted from Brazilian bankchecks is used to test the proposed system.