Comparison of Handprinted Digit Classifiers

We report recognition results for several pattern classifiers trained and tested on disjoint sets of 30620 digits selected from the first 500 writers of NIST Special Database 3. The classifiers are ubiquitous in traditional pattern recognition literature (minimum distance, maximum a posteriori, nearest neighbor) as well as neural network literature (multilayer perceptron, radial basis functions, probabilistic neural network). For the purpose of valid comparison of classifiers fixed sets of Karhunen-Loeve Transforms, were used as features. These were produced from images preprocessed using the fixed methods for size and orientation normalization. The “Kmeans” clustering algorithm is used to produce subclasses thereby supervising training and aiding recognition. Graphical displays of classification and associated confidences illustrate classifier complexity. Recognition error rates for all the classifiers are tabulated as a function of feature vector dimension. Computational and memory requirements of the different classifiers are also compared.

[1]  Patrick J. Grother,et al.  Cross-validation comparison of NIST OCR databases , 1993, Electronic Imaging.

[2]  Martin Fodslette Møller,et al.  A scaled conjugate gradient algorithm for fast supervised learning , 1993, Neural Networks.

[3]  Majid Ahmadi,et al.  Statistical and neural classification of handwritten numerals: a comparative study , 1992, Proceedings., 11th IAPR International Conference on Pattern Recognition. Vol.II. Conference B: Pattern Recognition Methodology and Systems.

[4]  Mohamad T. Musavi,et al.  On the training of radial basis function classifiers , 1992, Neural Networks.

[5]  Thomas G. Dietterich,et al.  Improving the Performance of Radial Basis Function Networks by Learning Center Locations , 1991, NIPS.

[6]  Majid Ahmadi,et al.  Pattern recognition with moment invariants: A comparative study and new results , 1991, Pattern Recognit..

[7]  Kunihiko Fukushima,et al.  Character recognition with selective attention , 1991, IJCNN-91-Seattle International Joint Conference on Neural Networks.

[8]  Belur V. Dasarathy,et al.  Nearest neighbor (NN) norms: NN pattern classification techniques , 1991 .

[9]  F. Girosi,et al.  Networks for approximation and learning , 1990, Proc. IEEE.

[10]  Wayne Ieee,et al.  Entropy Nets: From Decision Trees to Neural Networks , 1990 .

[11]  Donald F. Specht,et al.  Probabilistic neural networks , 1990, Neural Networks.

[12]  Lawrence D. Jackel,et al.  Handwritten Digit Recognition with a Back-Propagation Network , 1989, NIPS.

[13]  Teuvo Kohonen,et al.  Self-Organization and Associative Memory , 1988 .

[14]  Isabelle Guyon,et al.  Neural Network Recognizer for Hand-Written Zip Code Digits , 1988, NIPS.

[15]  Mehdi Hatamian,et al.  Optical character recognition by the method of moments , 1987 .

[16]  Geoffrey E. Hinton,et al.  Learning representations by back-propagating errors , 1986, Nature.

[17]  Ching Y. Suen,et al.  Automatic recognition of characters by Fourier descriptors and boundary line encodings , 1981, Pattern Recognit..

[18]  Robert S. Ledley,et al.  Special issue on Optical Character Recognition , 1970, Pattern Recognit..

[19]  R. Casey Moment normalization of handprinted characters , 1970 .

[20]  Geoffrey H. Ball,et al.  ISODATA, A NOVEL METHOD OF DATA ANALYSIS AND PATTERN CLASSIFICATION , 1965 .

[21]  Nils J. Nilsson,et al.  Learning Machines: Foundations of Trainable Pattern-Classifying Systems , 1965 .

[22]  C. M. Reeves,et al.  Function minimization by conjugate gradients , 1964, Comput. J..

[23]  Franz L. Alt,et al.  Digital Pattern Recognition by Moments , 1962, JACM.