Experimental analysis of the modified direction feature for cursive character recognition

This paper describes and analyzes the performance of a structural feature extraction technique for the recognition of segmented/cursive characters that may be used in the context of a segmentation-based, handwritten word recognition system. The modified direction feature (MDF) extraction technique builds upon a previous technique proposed by the authors that extracts direction information from the structure of character contours. This principle is extended so that the direction information is integrated with a technique for detecting transitions between background and foreground pixels in the character image. The MDF technique used in conjunction with neural network classifiers provide recognition rates of up to 90.24%, which are amongst the highest in the literature. This paper also presents a detailed analysis of the characters that were the source of misclassification in the character recognition process. The characters used for experimentation were obtained from the CEDAR benchmark database.