Chinese character recognition via Gegenbauer moments

Moment features have been applied in pattern recognition systems since the moment method was developed. In this paper, a set of moment features extracted from the newly developed Gegenbauer moment method for Chinese character recognition is proposed. Compared with results based on other moment methods, Gegenbauer moments can provide a modest improvement in terms of recognition for those Chinese characters that are very similar in shape. We conclude that the proposed new Gegenbauer moment method can supplement existing Chinese character recognition techniques based on local structure features.