Recognition of Handwritten Chinese Characters Based on Concept Learning

Abstract

Many deep-learning character recognition methods have been developed over the past few years. Chinese characters are widely used in many countries; however, the deep-learning-based Chinese character recognition methods are faced with various problems, such as a large amount of data required for training, numerous parameters, and a large consumption of computing resources. Concept learning is a hominine learning approach. Unlike existing deep-learning models, conceptual model learning can be realized by using as little as one sample. This paper is the first to propose a handwritten Chinese character recognition method based on concept learning. Different from the existing image representation-based character recognition methods, the proposed method builds a meta stroke library with prior knowledge, and then, presents a Chinese character conceptual model based on stroke relationship learning using a character stroke extraction method and Bayesian program learning. During character recognition, Monte Carlo Markov chain sampling is utilized to obtain the character generation model for each character conceptual. This generation model can calculate the probability of the target and training characters being the same classification, and thereby determines the classification of the target character. The experimental results indicate that, with the proposed method, the conceptual model of each character can be built for character classification prediction using as few as one character sample. Our approach obtains better performance than the state-of-the-art methods on ICDAR-2013 competition dataset. INDEX TERMS Character conceptual model, character recognition, concept learning, stroke extraction.

Recognition of Off-line Handwritten Uyghur Words Using Bayesian Networks with Grapheme Nodes

“…The segmentation-based recognition strategy can solve the problems that the holistic word recognition algorithm has poor discrimination ability and the algorithm's difficulty to expand. Attempts have been made on cursive text recognition based on the segmentation strategy [11][12][13][14][15]. For offline Arabic word recognition, Parvez et al [11] proposed a character segmentation and word recognition algorithm that combines structural features with a fuzzy polygon matching algorithm.…”

“…Some scholars have sought to find a more suitable segmentation unit than characters. For example, Xu Liang et al [14] built a meta-stroke library with prior knowledge and presented a Chinese character conceptual model based on stroke relationship learning. Partha Pratim Roy et al [15] decomposed text lines into character primitives and used string matching to spot words in historical documents.…”

“…To avoid unstable character segmentation results, some researchers have proposed using parts of characters, which were called graphemes or primitives, to replace the whole character as the segmentation unit [14,15]. Moreover, feature design for handwritten word image is another important factor that influences the performance of a handwriting recognition system, because extracting good features from a word image is difficult when the image is degraded or handwriting variability exists in the word.…”

Local Water-Filling Algorithm for Shadow Detection and Removal of Document Images

“…The shadows may make the perception of documents uncomfortable to the human eye and cause the degradation of text in documents or notes, which will result in difficulties for text binarization and recognition [12,13]. Therefore, removing shadows from document images not only helps generate clear and easy-to-read text [14], but also makes document binarization [15,16] and recognition tasks [17][18][19] possible.…”

Optimally configured convolutional neural network for Tamil Handwritten Character Recognition by improved lion optimization model

“…There have not been sufficient studies on TCR with machine learning models, but still, it is in the infant stage [15,35]. Among the countable researchers to recognize handwritten Tamil characters, the fuzzy concept is more primitive as it has the potential of string matching [30,41]. But, here the textual style of the test data was not mentioned at all [1].…”