Wavelet Packet Transform and Neuro-fuzzy Approach to Handwritten Character Recognition

This paper presents a novel method for automatic handwritten character recognition by combining wavelet packet transform with neurofuzzy approach. The time-frequency localization and compression capability of wavelet packet transform using best-basis algorithm is used for feature extraction, enhancing the accuracy of recognition at pixel level. The best-basis algorithm automatically adapts the transform to best match the characteristics of the signal, minimizing the additive cost function. Since fuzzy sets and fuzzy logic remains as a means for representing, manipulating and utilizing uncertain information and to provide a framework for handling uncertainties and imprecision associated with real world problems, a fuzzy logic system is used for classification purpose. A neural network system is used for recognition purposes since they provide computational power, fault tolerance, and learning capability to the systems. Characteristic features are extracted by taking wavelet packet transform using best-basis algorithm and are given as input to the fuzzy classifier where they are fuzzified and classified using IF ... THEN rules, and given to a neural network recognition system. This method is more efficient for handwritten character recognition as well as personal identification compared to energy sorted wavelet transform of character images, since characters contain very few edges in the images. Simulation of characters is done for 3 multiresolution levels using symmlet and results show that this method is more efficient than the methods using only fuzzy logic.