Invariant handwritten Chinese character recognition using fuzzy min-max neural networks

Chiu, Hung-Sheng; Tseng, Din-Chang

doi:10.1016/s0167-8655(97)00029-9

Cited by 37 publications

(10 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Ten numerals from one hundred writers are scanned and stored in BMP format. After moment normalization [10], the rotation invariant ring-data features defined by Ueda and Nakamura [11] and extended by Chiu and Tseng [12], are extracted from the character by setting ring width to two. The extracted ringdata vector is a 16-dimensional feature vector.…”

Section: ) Sonar Data Setmentioning

confidence: 99%

Fuzzy Hyperline Segment Neural Network Pattern Classifier with Different Distance Metrics

Kadam¹,

Bagal²

2014

IJCA

View full text Add to dashboard Cite

The Fuzzy Hyperline Segment Neural Network (FHLSNN) pattern classifier utilizes fuzzy set as pattern classes in which each fuzzy set is a union of fuzzy set hyperline segments. The Euclidean distance metric is used to compute the distances to decide the degree of membership function. In this paper, the use of other various distance metrics such as Manhattan, Squared Euclidean, Canberra and Chebyshew distance metrics is proposed. The performance of FHLSNN pattern classifier is evaluated with various benchmark databases such as Glass, Wine, PID, and Iris data set and real handwritten database. The FHLSNN pattern classifier is evaluated for generalization performance under recognition rate, training time and testing time. From the result analysis, the performance of classifier is based on the distance metrics as well the database used is verified. This analysis will help to select a suitable distance metric for fuzzy neural network classifier for particular application.

show abstract

Section: ) Sonar Data Setmentioning

confidence: 99%

Fuzzy Hyperline Segment Neural Network Pattern Classifier with Different Distance Metrics

Kadam¹,

Bagal²

2014

IJCA

View full text Add to dashboard Cite

show abstract

“…The back-propagation learning algorithm it used is time-consuming, since it takes a large number of iterations for the training data to converge to the stable state (Lippman, 1989;Wu et al, 1997). Chiu and Tseng (1997) used fuzzy min-max neural networks (Simpson, 1992) to classify the ring-data vectors that are extracted from thinned or nonthinned characters. The recognition rate is between 88% and 94% for single-writer characters, and is between 27% and 58% for multiwriter characters.…”

Section: Figure 24mentioning

confidence: 99%

A stroke-based neuro-fuzzy system for handwritten chinese character recognition

Lin¹,

Lee²,

Yang³

2001

Applied Artificial Intelligence

View full text Add to dashboard Cite

“…We have proposed new features called "generalized ring averaging", which is extension to ring features defined in [3]. Ring features, which are invariant to rotation and used for Chinese handwritten character recognition in [3] is applicable only to binary images. But it is redefined for any gray scale image for face recognition under change of direction of illumination.…”

Section: Introductionmentioning

confidence: 99%

Generalized Ring averaging: a new method for left and right directional illumination invariant face recognition for frontal poses and their small variants.

Potey¹,

Upasani²

2010

IJCA

View full text Add to dashboard Cite

In this paper a new method is proposed for face recognition when subject is illuminated from left and right direction in a fixed pose named "generalized ring averaging". We have proposed new features called "generalized ring averaging" features, which is extension to ring features. Ring features are invariant to rotation and used for binary images only. Proposed features are invariant to direction of illumination (left and right) and used for gray scale images. Well known Fuzzy min-max neural network classifier is used for recognition purpose. The proposed method is found better than one of the most popular method used for face recognition called "eigenfaces", in terms of percentage recognition rate, when compared with same dimensionality of feature vector. The proposed method requires less time to extract features than eigenfaces and recall time per pattern is found comparable to eigenfaces. However, the proposed method is suitable only for frontal poses and its little variants, which are very close to frontal pose and only for left and right direction of illumination by keeping pose and illumination strength around constant.

show abstract

Invariant handwritten Chinese character recognition using fuzzy min-max neural networks

Cited by 37 publications

References 26 publications

Fuzzy Hyperline Segment Neural Network Pattern Classifier with Different Distance Metrics

Fuzzy Hyperline Segment Neural Network Pattern Classifier with Different Distance Metrics

A stroke-based neuro-fuzzy system for handwritten chinese character recognition

Generalized Ring averaging: a new method for left and right directional illumination invariant face recognition for frontal poses and their small variants.

Contact Info

Product

Resources

About