Abstract—Moment descriptors have long been applied in object recognition since the early years of the development of the moment theories. Nowadays, discrete orthogonal moments have been studied and proposed for they are superior to traditional continuous ones. In this paper, a set of moment features extracted from the discrete Tchebichef moments for Chinese character recognition is presented. A new method of evaluating the variance values of each moment feature is applied in this research. Tested on a set of 6,763 Chinese characters, our newly proposed Tchebichef moment features perform very well in distinguishing all Chinese character pairs that have similar structures. Index Terms—Discrete orthogonal moments, tchebichef moments, Chinese character recognition. I. INTRODUCTION Since Hu [1] introduced the moment methods in 1961, moment descriptors have been widely used in image representation, pattern recognition, and object classification. One of the applications using moment features is the Chinese character recognition. Chinese characters are very different from many other languages in conveying information, while the structure of a character is a key to its meaning and pronunciation. Many feature extraction methods applied in the Chinese character recognition systems are based on the local features, such as strokes and feature points [2]-[5]. While the existing methods are quite efficient in general, there are some difficulties to distinguish two characters when they have very close structures. On the other hand, the moment method has the advantage of utilizing the global features of the Chinese characters, therefore, it can be used as a complementary scheme to overcome the obstacles confronted by other systems. Some Chinese character recognition systems based on the orthogonal moment descriptors have been reported [6]-[8]. Recently, the emergence of discrete orthogonal moments has substantially enriched the moment methods. Moments based on Tchebichef and Krawtchouk polynomials were introduced by Mukundan et al. [9] and Yap et al. [10] in 2001 and 2003, respectively. Since the computation of discrete orthogonal moments is not involved with numerical approximation of the weight function, while maintaining the property of orthogonal, the discrete moments are superior to traditional continuous orthogonal moments in terms of calculation accuracy and speed. Since they were introduced, the discrete orthogonal moments have been productively utilized in image processing [11]-[13], pattern recognition [14], [15], and many other scientific field [16]. In this research, we have utilized the Tchebichef moments in the aspect of Chinese character recognition. Four lower orders Tchebichef moments with the highest variance values are selected as features in the four dimensional Tchebichef moment space. A set of 6,763 Chinese characters, which are defined in the Chinese standard GB2312, is used as the testing characters. Our results show that the recognition ability of the four features based on the Tchebichef moments overwhelms the ones utilizing continuous orthogonal moments. II. TCHEBICHEF MOMENTS The definition of the n-th order classical Tchebichef Polynomial is defined as 3 2 () (1 ) ( , ,1 ;1,1 ;1) n n t x N F n x n N       (1) hypergeometric function. k a is the Pochhammer symbol given by         1 ... 1 k a k a aa a k a         (2)     2 1 0 ,;; ! k k k k k a b z F abcz c k     (3) With the definitions above, (1) can also be written as 0 1 () ! ( 1) n n k n k N k n k x t x n n k n k                            (4) The scaled Tchebichef polynomials are defined as () () (, ) n n t x t x nN   (5) where   n t x is the discrete Tchebichef polynomial given by (1), and   , nN  is a constant which is independent of x . Under the discussion above, the squared norm of the scaled polynomial is given by 2 (, ) (, ) (, ) nN pnN nN    (6) where   , nN  is Chinese Character Recognition by Tchebichef Moment Features Bing Hu and Simon Liao, Member, IACSIT Lecture Notes on Software Engineering, Vol. 1, No. 4, November 2013 392 DOI: 10.7763/LNSE.2013.V1.83 where   , 0,1, 2,..., , 0, 0,1 xn NN p    . 2 1 F is the Manuscript received June 12, 2013; revised July 30, 2013. Bing Hu is with the Applied Computer Science Department, University of Winnipeg, Manitoba, Canada (e-mail: hu-b86@webmail.uwinnipeg.ca). Simon Liao is with the Applied Computer Science Department, University of Winnipeg, Manitoba, Canada (e-mail: s.liao@.uwinnipeg.ca).