HMM-based recognition engine using a novel approach for statistical feature extraction

This paper extracts statistical features using a novel approach. The feature set locally measure the characteristics of the image. The proposed approach encodes the extracted features, from a one-pixel width window that slides horizontally the word image. We then inject the feature vector set into a recognition engine. The recognition engine is built using Hidden Markov Models Tool Kit (HTK). The system is trained and tested on the Arabic Printed Text Image (APTI) database. In order to select the optimal parameters for the HMM classifier, the APTI training dataset is further divided into a smaller training subset and a verification set. The estimated parameters are, then, used in the testing phase. The presented technique provides state-of-the-art recognition results on the APTI database using HMMs. The overall system achieved a recognition rate more than 97%.