Yasuji Miyake

Modified Quadratic Discriminant Functions and the Application to Chinese Character Recognition

Issues in the quadratic discriminant functions (QDF) are discussed and two types of modified quadratic disriminant functions (MQDF1, MQDF2) which are less sensitive to the estimation error of the covariance matrices are proposed. The MQDF1 is a function which employs a kind of a (pseudo) Bayesian estimate of the covariance matrix instead of the maximum likelihood estimate ordinarily used in the QDF. The MQDF2 is a variation of the MQDF1 to save the required computation time and storage. Two discriminant functions were applied to Chinese character recognition to evaluate their effectiveness, and remarkable improvement was observed in their performance.

Improvement of handwritten Japanese character recognition using weighted direction code histogram

Several algorithms for preprocessing, feature extraction, pre-classification, and main classification are experimentally compared to improve the recognition accuracy of handwritten Japanese character recognition. The compared algorithms are three types of nonlinear normalization for the preprocessing, the discriminant analysis and the principal component analysis for the feature extraction, the minimum distance classifiers and the linear classifier for the high-speed pre-classification, and modified Bayes classifier and subspace method for the robust main classification. The performance of the recognition algorithm is fully tested using the ETL9B character database. The recognition accuracy of 99.15% at the recognition speed of eight characters per second is achieved. This accuracy is the best one ever reported for the database.

Increasing the feature size in handwritten numeral recognition to improve accuracy

The relationship between the recognition rate of handwritten numerals and the normality of the distribution of their features has been investigated experimentally with a large amount of data in various combinations of quantized orientations and regions. The recognition method is based on the histogram of local orientation of contours of each numeral. To obtain a more accurate orientation quantization, the effectiveness of the orientation quantization using the gray-scale gradient has also been investigated. The results show that : (1) to increase the dimensionality of features, it is better to increase the number of quantized orientations, keeping the number of regions small (e.g., 4 x 4 or 5 x 5) ; (2) in the same dimensionality, the better the normality of a feature distribution, the higher the recognition rate ; (3) a quantization of orientations using gray scales is effective for normalizing a feature distribution ; and (4) the filter processing in reduction of the number of quantization scales improves the normality and recognition rate. The recognition of handwritten numerals collected from actual posts were carried out by using the gray-scale local-orientation histogram (400 dimensions). A correct recognition rate of 99.18 percent (mean value) has been obtained.

Handwritten numeral recognition using autoassociative neural networks

Describes the result of a fundamental study on pattern recognition using autoassociative neural networks, and experimental comparison on handwritten numeral recognition by conventional multi-layered neural network and statistical classification techniques. As the statistical classification techniques, the projection distance method and the nearest neighbor method are employed. The relationship between the projection distance method which is based on the K-L expansion and three layered autoassociative networks is discussed, and it is shown that the three and five layered autoassociative networks are superior to the projection distance method. In the handwritten numeral recognition experiment, a total of 44862 numeral samples collected by IPTP are used to evaluate and compare the recognition rates of the autoassociative networks, the mutual associative network, the nearest neighbor method, and the projection distance method. The five layered autoassociative networks achieved the highest recognition rate in the handwritten numeral recognition experiment. The result of experiment together with the fundamental study show that the autoassociative networks have such characteristics that: (1) class independent training makes the possibility of local convergence less than that of the mutual associative network, (2) the networks possess the higher ability of dimension reduction and interpolation than the nearest neighbor method (3) they yield less misclassification due to subspace sharing than the projection method, (4) the five layered autoassociative network can fit a curved hypersurface to a distribution of patterns.

Handwritten ZIP code recognition using lexicon free word recognition algorithm

The paper describes a new approach to ZIP code recognition using a word recognition algorithm, where a numeral string is recognized as a word. The paper also describes an end to end ZIP code recognition system consisting of tilt/slant correction, line segmentation, word segmentation, ZIP code location, as well as the ZIP code recognition. Evaluation tests are performed using address block image samples collected from United States mail pieces. The results of isolated numeral recognition, manually extracted ZIP code recognition, and end to end ZIP code recognition are presented to show and discuss the advantage of the word recognition based numeral string recognition.

Machine and human recognition of segmented characters from handwritten words

Handwritten character recognition by human readers, a statistical classifier, and a neural network is compared to know the required accuracy for handwritten word recognition. Sample characters extracted from postal address words on mail pieces collected by USPS were used to evaluate human and machine performance. Experimental results show that: 1) when the characters are segmented from words and are randomly presented, the accuracy of the machine recognition is comparable with the average human recognition accuracy, 2) the neural network employing the feature vector of size 64 outperforms the statistical classifier employing the same feature vector, and that 3) the statistical classifier employing the feature vector of size 400 achieves comparable recognition rate with the best human reader.

F-ratio Based Weighted Feature Extraction for Similar Shape Character Recognition

Recognition of handwritten similar shaped character is a difficult problem and in character recognition system most of the errors occur from similar shaped characters. In this paper we proposed a novel feature extraction technique to improve the recognition results of two similar shaped characters. The technique is based on F-ratio (Fisher Ratio), a statistical measure defined by the ratio to the between-class variance and within-class variance. F-ratio modifies the feature vector of two similar shape characters by weighting the feature elements. This weighting scheme enhances the feature elements that belongs to the distinguishable portions of the similar shaped characters and reduces the feature elements of the common portion of the characters, so that similar shaped characters can be identified easily. We considered pair of handwritten similar shape characters of different scripts like Arabic/Persian, Devnagari English, Bangla, Oriya, Tamil, Kannada, Telugu etc. and we noted that f-ratio based feature weighting shows better recognition results.

Accuracy improvement of slant estimation for handwritten words

Handwritten words are usually slant or italicized due to the mechanism of handwriting and the personality. In order to improve the accuracy of character segmentation and recognition, Kimura et al. (1993) proposed a chain code method for the slant estimation and correction. However the method is very simple and usually gives good estimate of the word slant, there was a problem such that the slant tends to be underestimated when the absolute of the slant is close or greater than 45/spl deg/. To solve the problem, we proposed an iterative chain code method. We introduce a noniterative method using 8-directional chain code for improving the linearity and the accuracy of the slant estimation. The experimental results show that the proposed method improves the linearity and the accuracy of the slant estimation efficiently without sacrificing the processing speed and the simplicity.

On feature extraction for limited class problem

The availability of the canonical discriminant analysis to a limited class problem is restricted because the number of extracted features can not be or exceed the number of classes. In order to remove the restriction, a new feature extraction technique FKL is proposed and is tested by handwritten numeral recognition experiment. While the canonical discriminant analysis maximizes the variance ratio (F-ratio), and the principal component analysis (K-L expansion) minimizes the mean square error of dimension reduction, the FKL optimizes both the F-ratio and the mean square error simultaneously. The result of experiment shows that the FKL provides the richest features in discriminating power for the limited class problem when compared with other techniques including the canonical discriminant analysis, the principal component analysis, and the orthonormal discriminant vector method (ODV).

Evaluation and improvement of slant estimation for handwritten words

Handwritten words are usually slanted or italicized due to the mechanism of handwriting and ones personality. The authors have proposed a chain code method for slant estimation and correction to improve the accuracy of character segmentation and recognition. However, the method is very simple and usually gives a good estimate of the word slant but the relationship between the actual slant and the estimated slant is nonlinear, and the slant tends to be underestimated when the absolute of the slant is close or greater than 45/spl deg/. To resolve this, the paper proposes new methods for evaluating and improving the linearity and accuracy of slant estimation. The result of experimental study shows that the proposed algorithms are improved in the linearity and the accuracy of the slant estimation, and that a high speed iterative method is also efficient as far as processing time is concerned.