Structure Of Chinese Characters Research Articles

Rule-based fine-grained vocabulary for enhanced Chinese pre-training

Abstract Chinese pre-training is an essential direction in Chinese natural language processing, and vocabulary constitutes the foundation of pre-trained models. Existing methods for constructing vocabularies for Chinese pre-trained models typically treat each Chinese character as an indivisible token, overlooking the additional information embedded in the intrinsic structure of Chinese characters and its impact on model performance. To leverage this information, we propose a method of training with rule-based fine-grained vocabularies to directly learn the sequence of intrinsic structures of Chinese characters in Chinese pre-training. Specifically, we first construct a mapping rule-based fine-grained vocabulary based on the glyph and radical splitting mapping relationship of Chinese characters. Subsequently, we employ a whole char masking strategy to train Chinese pre-trained models based on this new vocabulary. Experimental results demonstrate that compared to the BERT baseline model, our model achieves promising performance on multiple downstream tasks, with fewer model parameters and stronger robustness. The proposed Chinese pre-training method exploits the intrinsic structural information of Chinese characters, providing a novel method for subsequent research in Chinese natural language processing.

Mythologizing of characters: Chinese character archaeology evidence of ancient Chinese totems related to “Yi (一)” cognate

This article endeavors to elucidate the mechanisms by which vestiges of mythology become embedded in the morphology of Chinese characters. It represents a foundational study aimed at explicating how oral mythological traditions and archaic mythical cognitions manifest themselves through the concrete graphemic structures of Chinese characters. The research focuses on the analysis of the characters “Yi (一),” “Yi (壹),” “Yi (羽/能),” and “Yi (鳦),” which are cognates. Through this examination, the study demonstrates that these graphemes are reflective of ancient Chinese veneration of the gourd, as well as bear and bird totemic traditions. To substantiate these claims, the methodology of “Chinese character archaeology” is employed, which integrates etymological analysis of the characters themselves with traditional textual sources, archaeological findings, and anthropological data. This approach can be extended to encompass a diverse array of totems, including plants, dragons, tigers, and horses. The ancient Chinese characters that encapsulate these totemic traditions are of particular significance as they offer insights into primordial modes of thought that predate the genesis of philosophical literature. Furthermore, this research potentially elucidates, through graphemics, the hypothesis that Chinese ancient civilization originated simultaneously in various regions with distinct totemic traditions, gradually converging into a predominant civilization.

Fine Segmentation of Chinese Character Strokes Based on Coordinate Awareness and Enhanced BiFPN.

Considering the complex structure of Chinese characters, particularly the connections and intersections between strokes, there are challenges in low accuracy of Chinese character stroke extraction and recognition, as well as unclear segmentation. This study builds upon the YOLOv8n-seg model to propose the YOLOv8n-seg-CAA-BiFPN Chinese character stroke fine segmentation model. The proposed Coordinate-Aware Attention mechanism (CAA) divides the backbone network input feature map into four parts, applying different weights for horizontal, vertical, and channel attention to compute and fuse key information, thus capturing the contextual regularity of closely arranged stroke positions. The network's neck integrates an enhanced weighted bi-directional feature pyramid network (BiFPN), enhancing the fusion effect for features of strokes of various sizes. The Shape-IoU loss function is adopted in place of the traditional CIoU loss function, focusing on the shape and scale of stroke bounding boxes to optimize the bounding box regression process. Finally, the Grad-CAM++ technique is used to generate heatmaps of segmentation predictions, facilitating the visualization of effective features and a deeper understanding of the model's focus areas. Trained and tested on the public Chinese character stroke datasets CCSE-Kai and CCSE-HW, the model achieves an average accuracy of 84.71%, an average recall rate of 83.65%, and a mean average precision of 80.11%. Compared to the original YOLOv8n-seg and existing mainstream segmentation models like SegFormer, BiSeNetV2, and Mask R-CNN, the average accuracy improved by 3.50%, 4.35%, 10.56%, and 22.05%, respectively; the average recall rates improved by 4.42%, 9.32%, 15.64%, and 24.92%, respectively; and the mean average precision improved by 3.11%, 4.15%, 8.02%, and 19.33%, respectively. The results demonstrate that the YOLOv8n-seg-CAA-BiFPN network can accurately achieve Chinese character stroke segmentation.

A Novel Multi-Stage Recognition System in Handwritten Chinese Characters

Due to the various individual handwriting styles and diverse structures of Chinese characters, it is rather intricate to recognize handwritten Chinese characters precisely. In this paper, a muti-stage system that combines a Differentiable Binarization Network (DBNet) for text detection and Scene Text Recognition with a Single Visual Model (SVTR) for effective character recognition is developed, aiming to offer a potential solution for correct identification of handwritten Chinese characters in real-world scenarios. More specifically, the proposed system utilized the CASIA-HWDB2.x dataset from the Institute of Automation of the Chinese Academy of Sciences to train text detection and recognition models. For the detection, the Differentiable Binarization method in the Head Network applies a dynamic threshold map for text-background differentiation. On top of the SVTR employed as a recognition module, this research proposed a performance-optimized PP-LCNet architecture and the Guided Training of Continuous Time-Series Classifiers (GTC) with TextConAug, which is a data augmentation strategy to accommodate real-world scenarios. Briefly, DBNet and SVTR are integrated to form an end-to-end system. The output of DBNet after undergoing cropping and direction classifier, guides SVTR for a comprehensive system, ensuring seamless detection and recognition of handwritten Chinese characters. Experimental results indicate that this system achieves a promising performance and can provide accurate information on handwritten Chinese characters.

Декомпозиционная структура китайского иероглифа

The aim of the study is to systematically describe the decomposition structure of Chinese characters. The article contains a structural and categorical analysis of Chinese writing units. The scientific novelty of the research consists in the formulation of such a typological characteristic of Chinese character writing as lists, i.e., the constant need to compile various lists of characters and their components. These are lists of strokes; lists of characters; lists of indexing components; lists of indexing components series including lists of characters they form; lists of commonly used characters; lists of characters for learning Chinese as a foreign language, etc. As a result of analysing Russian and foreign scientific works on the research topic and the current language standards developed by the State Language and Writing Work Committee of the PRC (国家语言文字工作委员会), concepts and terms on the research topic (“stroke”, “character component”, “radical”, “indexing component”, “determinative”, “phonetic”) were clarified and systematised; the classification of character components depending on their location, the correlation between the concepts of the indexing component and the character component in diachrony and the structure of a hieroglyph in the process of its decomposition, i.e., in the process of breaking a hieroglyph into its constituent elements, were clarified.

The Cognition of Time Shaped by Linguistic Elements: Alphabetic Language and Ideographic Language

This study investigated the impact of language script on spatial cognition by comparing reaction times of native speakers of alphabetic languages and Chinese in response to horizontally and vertically arranged button groups. It is postulated that the linear, one-dimensional nature of alphabetic scripts may shape cognitive processing differently than the complex, two-dimensional Chinese characters. In the experiment, participants from both linguistic backgrounds were tasked with quickly responding to buttons arranged in both horizontal and vertical configurations. Results indicated that speakers of alphabetic languages exhibited significantly shorter reaction times for the horizontally oriented buttons compared to the Chinese speakers. Conversely, Chinese speakers demonstrated faster reaction times for vertically arranged buttons relative to the alphabetic language speakers. These results suggest that language script structure can influence spatial cognition. The linear, sequential arrangement inherent to alphabetic languages may predispose speakers to be more efficient in processing horizontally sequenced stimuli, while the two-dimensional structure of Chinese characters may confer an advantage in processing vertically presented information. Additionally, cultural factors, such as the traditional vertical writing and reading direction of Chinese script, may contribute to these differences. The findings provide preliminary support for the notion that language script can shape cognitive strategies for processing spatial information, with speakers of alphabetic languages and Chinese employing different strategies for horizontal and vertical spatial processing. Further research is needed to explore the underlying mechanisms, as well as the impact of early exposure to a specific script on spatial cognition.

Generating Structurally Complete Stylish Chinese Font Based on Semi-Supervised Model

Stylish fonts are widely utilized in social networks, driving research interest in stylish Chinese font generation. Most existing methods for generating Chinese characters rely on GAN-based deep learning approaches. However, persistent issues of mode collapse and missing structure in CycleGAN-generated results make font generation a formidable task. This paper introduces a unique semi-supervised model specifically designed for generating stylish Chinese fonts. By incorporating a small amount of paired data and stroke encoding as information into CycleGAN, our model effectively captures both the global structural features and local characteristics of Chinese characters. Additionally, an attention module has been added and refined to improve the connection between strokes and character structures. To enhance the model’s understanding of the global structure of Chinese characters and stroke encoding reconstruction, two loss functions are included. The integration of these components successfully alleviates mode collapse and structural errors. To assess the effectiveness of our approach, comprehensive visual and quantitative assessments are conducted, comparing our method with benchmark approaches on six diverse datasets. The results clearly demonstrate the superior performance of our method in generating stylish Chinese fonts.

Design of a Dual-Band WiFi Antenna Using the Theory of Characteristic Modes and Nested Chinese Characters

In this paper, a dual-band WiFi antenna and its Multiple Input Multiple Output (MIMO) system application is designed, fabricated, and measured based on the Chinese characters “Men” and “Wei”. The antenna uses a 40 × 40 × 1.6 mm3 FR4 substrate to analyze the combinatorial structure of Chinese characters using the theory of characteristic modes, to optimize the antenna dimensions by analyzing the mode current distribution, and to broaden the antenna bandwidth by etching rectangular slots on the ground. The measured and simulated results show that the four-element MIMO antenna covers 5.68–8.01 GHz, the isolation between the antennas is higher than 20 dB in the working band, the envelope correlation coefficient and the channel capacity losses of the simulation are lower than 0.001 and 0.18 bits/s/HZ, respectively. The efficiency of the antenna is higher than 90%, and it can be used for WiFi communication bands (5.8 GHz and 6 GHz).

Metric and Color Modifications for the Automated Construction of Map Symbols

Personalized mappings become popular among the public with the support of data diversity and device diversity. To develop personalized maps, constructing map symbols through automated ways is beneficial. The formal representation of map symbols (i.e., expressing map symbols by mathematical operators) is fundamental to the automated construction of map symbols. A previous study to evaluate the feasibility of structures of Chinese characters for representing map symbols shows that 77.5% of map symbols can be represented by them, although there are imperfections in some cases. It means that: (1) the other 22.5% of symbols should be formally represented by other mathematical solutions, and (2) those imperfect cases should be made perfect through some modification or refinements. In this study, we solve the representation problems of these two types of map symbols (i.e., the map symbol did not or imperfectly fit the structures of Chinese characters) by employing additional basic operators and proposing some metric and color modifications. To validate these proposed solutions, experiments have been carried out by using eight sets of symbols that are publicly available (e.g., Google Icons). The results indicated that almost all the map symbols can be formally represented with additional operators and metric and color modifications. The percentages of map symbols that did not fit structures of Chinese characters solved by these operators and modifications are 2.4% and 20.1%, respectively. The percentages of map symbols that imperfectly fit them solved by these operators and modifications are 8.7% and 8%, respectively. This work could not only enrich cartographic theory but also prompt the mathematization of map symbol construction.

Visual analysis of inscriptions in the Tang Dynasty: a case study on the calligraphy style of Wang Xizhi

The art of inscriptions is one of the most important components of the glorious ancient culture in China. It is an essential method not only for ancients to record history and disseminate culture but also to provide original written and pictorial records for subsequent generations to learn about ancient Chinese culture. However, the art of inscriptions and calligraphy is currently not widely available because of its professional chirography, style and composition. Moreover, traditional calligraphy exhibitions cannot provide the public with an interactive and intuitive way to present the features of Chinese characters, such as stroke thickness. Therefore, in this paper, we present an interactive visual system to support the public in understanding and appreciating the calligraphic style of the inscriptions of the Tang Dynasty and the evolutionary path of the calligraphic style of Wang Xizhi. We first employ image processing technology to extract calligraphy features. Then, we help users explore the development of calligraphy from the spatial-temporal dimension and analyze the similarity between the works at two granularity levels: Chinese character structure and the works’ style. Furthermore, the system also provides a metaphorical visualization method to enhance the concretization of calligraphic appreciation. Case studies and comprehensive evaluation demonstrate the usability of our proposed visual analysis system of the calligraphic style of the inscriptions in the Tang Dynasty.

Research on the Formation of Chinese Characters

Alphabetical writing systems use letters that are closely related to speech to record speech. However, the glyphs of Chinese characters do not directly express sounds, and the ideographic mechanism between the glyphs and their meanings is not transparent. We analysed the relationship between the glyphs and meanings of nine thousand Chinese characters and found that the combination of elements in the glyph constructs certain imagery, and this imagery is a psychological representation of the experience informing the concept of the character, so the meaning and glyph of the characters are ideographically connected. In other words, the formation of Chinese characters is to use a small number of basic components with images to combine with each other to construct different imageries, so that the imagery can metaphor the meaning of the Chinese characters, including polysemy. In order to search for these imageries scientifically we propose four necessary conditions that the imageries sought must meet. We use the many related meanings derived from the configuration to infer the imagery of the configuration, so that this imagery can explain the meanings of the characters. In the end, we found out the imagery and structure these nine thousand Chinese characters, summarized the five combination modes of Chinese characters, and extended these combination modes to multi-character words combining characters and characters. These results reveal the ideographic mechanism of the hierarchical combination of the internal structure of Chinese characters and their imagery, unearth the interpretable imagery and knowledge that directly represent the meaning of Chinese characters, and help propose a guiding theory for character word formation in the Chinese writing system.

A Tree-Structure Analysis Network on Handwritten Chinese Character Error Correction

Existing researches on handwritten Chinese characters are mainly based on recognition network designed to solve the complex structure and numerous amount characteristics of Chinese characters. In this paper, we investigate Chinese characters from the perspective of error correction, which is to diagnose a handwritten character to be right or wrong and provide a feedback on error analysis. For this handwritten Chinese character error correction task, we define a benchmark by unifying both the evaluation metrics and data splits for the first time. Then we design a diagnosis system that includes decomposition, judgement and correction stages. Specifically, a novel tree-structure analysis network (TAN) is proposed to model a Chinese character as a tree layout. Using the predicted tree layout for judgement, if the character is wrongly written, correction operation is needed for error analysis. The correction stage mainly consists of three steps: fetch the ideal character, correct the errors and locate the errors. Computing the distance between the model output and all candidate characters can generate ideal characters. With the decomposition ability of TAN, edit distance computing between predicted tree and ideal tree can lead to correction operations. By using attention visualization, we can even provide the approximate error location. Through quantitative analysis, we prove that TAN captures more accurate spatial position information than regular encoder-decoder models. Additionally, we propose a novel bucketing mining strategy to apply triplet loss at radical level to alleviate feature dispersion. Experiments on handwritten character dataset demonstrate that our proposed TAN shows great superiority on all three metrics comparing with other state-of-the-art models.

Visual motion processing in Chinese children with developmental dyslexia: An fMRI study.

Dorsal stream is an important pathway for visual information transmission. As a part of the dorsal pathway, the middle temporal visual motion areas (V5/MT+) are mainly responsible for visual motion processing and the ability of visual motion processing is closely related to reading. Compared with alphabetic scripts, the visual structure of Chinese characters is more complex and there are no clear grapheme-phoneme correspondence rules. So the ability of visual analysis plays an important role in Chinese character processing. This study first investigated the brain activation of Chinese dyslexic children and children of the same chronological age when they observed coherent motion stimuli. ROI analysis indicated that only the activation of left V5/MT+ was significantly weaker in dyslexics than that in the control group. The activity of the magnocellular-dorsal stream was closely related to orthographic awareness in the combined data (two groups) and the typical children. In dyslexia group, the stronger the activation of V5/MT+ was, the worse the phonological awareness, rapid naming performance and orthographic awareness were. In short, Chinese dyslexic children were deficient in the activation of the left V5/MT+ and the activity of the magnocellular-dorsal pathway was closely related to orthographic awareness in Chinese pupils.

Chinese Character Components Segmentation Method Based on Faster RCNN

To solve the component segmentation problem caused by the sticking and overlapping of parts in incoherent handwritten calligraphy characters, we propose a Chinese character part segmentation method based on Faster RCNN. The method utilizes the advantages of Faster RCNN on multi-scale and small targets to solve difficult problems in component segmentation. The hierarchical features of the components were used in our proposed method to identify each layer of the Chinese character structure to obtain the components. Qualitative and quantitative calculations were used to test the segmentation effect of the proposed method. The experimental results demonstrate the accurate segmentation effectiveness of our method for adhering and overlapping components. In addition, these components could be retrieved accurately in the retrieval system, and the mean Average Precision of the top 30 retrieval results reached 95.7%. A better retrieval accuracy reflects a better segmentation effect from the side, which proves the effectiveness of the proposed method.

韩国中级水平汉语学习者汉字书写偏误与其汉字习得

From the perspective of Korean Chinese learners, this paper determines the unit of analysis for Chinese writing errors through investigation and analysis. On this basis, it makes use of corpus to analyze Chinese writing errors of learners, and then studies their Chinese character acquisition. Through the research, it is found that the Acquisition of Chinese characters by Korean intermediate Level Chinese learners is the whole character acquisition, the whole character is decomposed into various components （components, radicals）, according to the structural characteristics of each component for memory, among which the pronoun cue plays a significant role. Similar words in the form are the biggest trouble for learners at this stage, but the error rate of similar words in the sound is higher. Korean learners at this stage have strong ability to analyze the structure of Chinese characters and to comprehensively use the form, sound and meaning of Chinese characters. Their writing errors are mainly stroke deletion, substitution and dislocation. Finally, this paper discusses how to judge whether Chinese characters are acquired or not, and proposes that Chinese characters writing errors should be studied from the perspective of Chinese language and the relationship between Chinese characters and Chinese language.

Pronunciation-Enhanced Chinese Word Embedding

Chinese word embeddings have recently garnered considerable attention. Chinese characters and their sub-character components, which contain rich semantic information, are incorporated to learn Chinese word embeddings. Chinese characters can represent a combination of meaning, structure, and pronunciation. However, existing embedding learning methods focus on the structure and meaning of Chinese characters. In this study, we aim to develop an embedding learning method that can make complete use of the information represented by Chinese characters, including phonology, morphology, and semantics. Specifically, we propose a pronunciation-enhanced Chinese word embedding learning method, where the pronunciations of context characters and target characters are simultaneously encoded into the embeddings. Evaluation of word similarity, word analogy reasoning, text classification, and sentiment analysis validate the effectiveness of our proposed method.

Ancient Chinese Character Image Retrieval Based on Dual Hesitant Fuzzy Sets

The complex and changeable structures of ancient Chinese characters result in the decreasing accuracy of their image retrieval. To resolve this problem, a new retrieval method based on dual hesitant fuzzy sets is proposed. Dual hesitation fuzzy sets that can express uncertain information more comprehensively are employed in the feature extraction process of directional line elements. The multiattribute evaluation index of adjacent grids for the current grid and its corresponding membership and nonmembership functions are established, and the weight of each attribute is calculated by the dual hesitation fuzzy entropy, such that the proposed features can fully reflect the topological structure of ancient Chinese characters. Using the dual hesitation fuzzy correlation coefficient to measure the similarity between the ancient Chinese character images to be retrieved and the candidate images, the retrieval of ancient Chinese character images is realized. Experiments show that when the t0hreshold value of the correlation coefficient is 0.9, the average retrieval accuracy is 90.4%.

FSPRM: A Feature Subsequence Based Probability Representation Model for Chinese Word Embedding

Chinese word embedding models capture Chinese semantics based on the character feature of Chinese words and the internal features of Chinese characters such as radical, component, stroke, structure and pinyin. However, some features are overlapping and most methods do not consider their relevance. Meanwhile, they express words as point vectors that cannot better capture different aspect semantics of Chinese words. In this paper, we propose a Feature Subsequence based Probability Representation Model (FSPRM) for learning Chinese word embeddings, in which we first integrate the morphological and phonetic features (stroke, structure and pinyin) of Chinese characters and learn their relevance by designing a feature subsequence to capture relatively comprehensive semantics of Chinese words, then feature probability distribution is proposed for capturing different aspect meanings of Chinese words based on the three internal features and probability representation by estimating its mean as the sum of feature subsequences. Chinese words with similar features may have similar semantics, then we map Chinese words to feature probability distributions and design a similarity-based objective for predicting the contextual words of the target word to learn their semantics. Extensive experiments on word analogy, word similarity, text classification and named entity recognition tasks demonstrate that the proposed method outperforms most state-of-the-art approaches.

ИЕРОГЛИФ И ИКЕБАНА: ОТ БАЛАНСА ИНЬ-ЯН К РАЗРУШЕНИЮ ФОРМЫ СИН-ГЁ-СО

This study investigates a deep connection between Japanese characters and Ikebana, which is due to the fact that Japanese characters as well as ikebana express the same picture of the world. Because the basis of this model of the world is laying in the Chinese tradition, we will first examine the Chinese characters and Chinese art of flower arrangement or chahua, and then the Japanese characters and Japanese ikebana. In the study, we refer not only to the works of Russian researchers, such as V.V. Malyavin, V.G. Belozyorova and V.A. Pronnikov, but, to a greater extent, to sources in Japanese (Nakamura Yasuo, Sasaoki Ryuho, Kumakura Isao, etc.), as well as to English-speaking authors or translations (Li Xia, Li Huilin). Analyzing the structure of the Chinese character, we look at lined paper for calligraphy which is usually being used in China. Lines of this paper indicate the connection of the structure of a character with Taiji or Supreme Ultimate, because a character is inscribed in a circle of Taiji, it has Yin and Yang sides and as we can say ‘exists’ in the ‘circle of changes’ or Bagua. In Japan, Chinese lined paper were not widespread, but the styles of calligraphy (regular script, semi-cursive script, cursive script) that came from China were elevated to the principle of simplified and broken form, Shin – Gyo – So, which became the basis for the Japanese type of beauty, kuzushi-no bi. Then we turn to the Chinese flower art, chahua and show its basic principle, which related to the Chinese characters, we mean the balance of Yin and Yang. Japanese art of ikebana we analyze in more detail and observe three main schools of Ikebana: Ikenobo, Ohara and Sogetsu. In the schemes of the Ikenobo (the oldest school in Japan), we find schemes in which floral arrangements are inscribed in a circle of Taiji (rikka flowers). In a later direction of Japanese flower arrangement (shoka, or seika flowers), we meet the principle of Shin – Gyo – So. We consider the same principle on the example of chabana, that is, flower arrangements for a tea ceremony. We also observe ikebana of Japanese literati Bunjin, who strive for a lifestyle of Chinese literati Wenren. The flowers in Chinese taste in contrast with actually Japanese ikebana do not have any patterns, but it has ‘refined riddle theme’ meigogadai with limited combinations of flowers. Also, it has more obvious connection with Chinese characters form. We conclude that a Chinese character shows subtle difference between manifested and unmanifested world, balance of Yin Yang while in Japanese tradition we can see preference of forms broken or simplified with Shin Gyo So. Chinese flower arrangement сhahua is also based on Yin Yang balance while in Japanese Ikebana it was transformed with Shin Gyo So.

Separating Chinese Character from Noisy Background Using GAN

Separating printed or handwritten characters from a noisy background is valuable for many applications including test paper autoscoring. The complex structure of Chinese characters makes it difficult to obtain the goal because of easy loss of fine details and overall structure in reconstructed characters. This paper proposes a method for separating Chinese characters based on generative adversarial network (GAN). We used ESRGAN as the basic network structure and applied dilated convolution and a novel loss function that improve the quality of reconstructed characters. Four popular Chinese fonts (Hei, Song, Kai, and Imitation Song) on real data collection were tested, and the proposed design was compared with other semantic segmentation approaches. The experimental results showed that the proposed method effectively separates Chinese characters from noisy background. In particular, our methods achieve better results in terms of Intersection over Union (IoU) and optical character recognition (OCR) accuracy.