Character Font Research Articles

Aesthetic Ideas in Chinese Calligraphy and Characters

Chinese aesthetic thought is rooted in the deep cultural heritage of China, which has been inherited and developed for thousands of years. Aesthetic thought has influenced the development of traditional Chinese art and culture, and also shaped and developed Chinese characters. As a system of symbols for recording Chinese language and passing on culture, Chinese characters have taken on forms that reflect the imprint of Chinese aesthetic thought on Chinese characters. Based on the development of Chinese characters and the evolution of the art of calligraphy, and through the elaboration of the characters hao, zhong and ren, This paper sheds light on the connection between the development of Chinese characters and the art of calligraphy, and how Chinese aesthetic thought contributes to the growth of both. This study reveals how traditional Chinese aesthetics shape the aesthetic form of square characters by analyzing the evolution of Chinese character fonts and traditional Chinese aesthetics. Moreover, calligraphy, as an artistic expression of Chinese character writing, is deeply influenced by aesthetic ideas in its unique writing style and chapter layout.

Rejecting small! Accepting larger? Optimizing the font size of Chinese characters basing legibility and visual fatigue in OST-HMDs

Font size is highly related to the legibility and visual fatigue in OST-HMDs, but the effects of font size on these factors remain further explored. An experiment was conducted to investigate the effects of a wider range of Chinese character font size (0.32°–1°) on legibility and visual fatigue, as well as to determine the optimal font size. Results showed that 0.32° had the worst legibility, but there was no continuous improvement as font size increased. A larger font size was found to be beneficial in reducing visual fatigue until it reached 0.95°, beyond which visual fatigue would relatively increase. Font size smaller than 0.32° should be rejected while a larger font size does not always provide more benefits. Considering legibility, visual fatigue and efficiency of text presentation, 0.84° is a relatively optimal Chinese character font size.

Design and Implementation of Dongba Character Font Style Transfer Model Based on AFGAN.

Dongba characters are ancient ideographic scripts with abstract expressions that differ greatly from modern Chinese characters; directly applying existing methods cannot achieve the font style transfer of Dongba characters. This paper proposes an Attention-based Font style transfer Generative Adversarial Network (AFGAN) method. Based on the characteristics of Dongba character images, two core modules are set up in the proposed AFGAN, namely void constraint and font stroke constraint. In addition, in order to enhance the feature learning ability of the network and improve the style transfer effect, the Convolutional Block Attention Module (CBAM) mechanism is added in the down-sampling stage to help the network better adapt to input font images with different styles. The quantitative and qualitative analyses of the generated font and the real font were conducted by consulting with professional artists based on the newly built small seal script, slender gold script, and Dongba character dataset, and the styles of the small seal script and slender gold script were transferred to Dongba characters. The results indicate that the proposed AFGAN method has advantages in evaluation indexes and visual quality compared to existing networks. At the same time, this method can effectively learn the style features of small seal script and slender gold script, and transfer them to Dongba characters, indicating the effectiveness of this method.

Review of GAN-Based Research on Chinese Character Font Generation

Review of GAN-Based Research on Chinese Character Font Generation

A Robust Approach for Licence Plate Detection Using Deep Learning

Intelligent transport systems must be developed due to the rising use of vehicles, particularly cars. In the field of computer vision, the identification of a vehicle's licence plate (LP) has been crucial. Various methods and algorithms have been used for the detection process. It becomes challenging to find similar photos, nevertheless, because the features of these plates change depending on colour, font, and language of characters. The research proposes a powerful deep learning framework based on feature extraction using convolutional neural networks and localization using canny-edge detection. Three steps make up the model's operation. An improved approach integrating the usage of bilateral filters and Canny edge detection is used for the processes of segmentation and localization. Further, a CNN architecture is used to extract features from images and classify the presence of licence plates in unseen vehicles. If present, the stage is followed by recognition of numbers written on the plates. An extensive experimental investigation takes place using three datasets namely Stanford Cars, Car Licence Plate Detection dataset and Indian Licence Plates Database. The attained simulation outcome ensures a superior performance over existing techniques in a significant way.

SMFNet: One-Shot Recognition of Chinese Character Font Based on Siamese Metric Model

SMFNet: One-Shot Recognition of Chinese Character Font Based on Siamese Metric Model

Deciphering broken Chinese characters on a broken piece (#10447) of Huayanshijing stored in Korea

This study aimed to decipher the broken Chinese characters on a fragment of Huayanshijing, #10447. All the broken characters have lost parts of the whole character fonts, and many have been severely damaged. These have made it difficult to decipher them to determine the original forms. This study deciphered the original forms by comparing the font of the broken Chinese characters with characters of similar forms from the Huayanshijing text and applying them to the sentences containing the characters in Huayanshijing.

書畫 同體·同出·同源論 形成의 哲學的 背景

This article is to reveal the meaning of the thoughts of SeoHwaDongChe (書畵同體: Calligraphy and drawing are same as each other), SeoHwaDongChul (書畵同出: Calligraphy and drawing are beginning together) and SeoHwaDongWon (書畵同源: Calligraphy and drawing sprout from same root) focusing on the understanding of the correlation between Seo (Calligraphy) and Hwa (Painting) that appeared on the stream of the aesthetics of Seo and Hwa flowing through such eras as Tang, Song and Yuan dynasties in China. The transition process of the relation between Seo and Hwa from DongChe to DongChul that finally went to DongWon embraces such art-philosophical leveled question as ‘Where did Seo and Hwa come from’, which shows the art spirit and aesthetic goal of the literary noblemen who approached the creation of Seo and Hwa. Such thoughts also contain the idea that painting works must be expressed with YongPil (用筆: ways of brushing) and PilEui (筆意: intention of brushing) of calligraphy in order to fully realize the art world sought for by the literary noblemen. First of all, in this aspect, it is required to analyze what kind of art calligraphy is in aesthetic and philosophical viewpoints in order to correctly understand the thoughts of DongChe, DongChul, DongWon and DongBeop (書畵: Calligraphy and painting are same in rules of creating works). In the Chinese history of SeoHwa (Calligraphy and Painting), creation of every piece of work begins from drawing a stroke. The commonness between Seo and Hwa and SeoHwaHyoNeungIlCheSeol (書畫效能一體說: the theory that Seo and Hwa have same effect) argued by Shí Tāo (石濤) based on his IlHoikRon (一劃論: the theory of one stroke) emphasized that every painter must learn the calligraphic principle for actual art creation activity. Looking into the relation between calligraphy and painting with consideration of such emphasis, it is desirable to create painting pieces using PilHoik (brush and stroke) of calligraphy. That is, the thoughts of DongChe, DongChul, DongWon and DongBeop are thought to be as same as the philosophical thoughts related with use of one stroke. In the tendency of art creation, calligraphy has CheeDoSeol (取道說: the theory of taking truth) leveled art spirit with which the artist must take the principle of truth in metaphysical aspect and set such truth as the base of calligraphy. Such CheeDoSeol based calligraphic forms are related with the thought of GwanMulCheeSangRon (觀物取象論: the theory of taking forms by viewing things), which causes artists to give painting leveled meaning to the character forms and fonts simply showing why painting must take PilBeop (rules of brushing) and PilEui of calligraphy. In such aspects set forth above, therefore, this article looks into how the theories of Seo and Hwa have been established and transited in relation with the perception of calligraphy and painting in the Chinese history of Seo and Hwa in the viewpoints of IlHoikRon, SeoHwaHyoNeungIlCheSeol, CheeDoSeol, metaphysical perception of calligraphy, GwanMulCheeSangRon and painting styled character fonts and forms.

Real-time high quality font generation with Conditional Font GAN

Designing and generating novels fonts manually is a laborious and time-consuming process owing to the large number and complexity of characters in the majority of language systems. Recent advancements in generative adversarial networks (GANs) have significantly improved font generation. These GAN-based approaches either handle the font generation as a vanilla GAN problem (that is, by synthesizing characters from a uniform latent vector) or an image-to-image translation problem. While the former approach has no limitation in generating diverse font styles, the generated fonts contain artifacts and can operate only on low-resolution images, thus impairing their usability. The latter approach generates high-quality font images for previously observed fonts, but the quality degrades during the inference phase while designing novel fonts. Furthermore, additional fine-tuning steps are required to achieve photorealistic results, which is computationally expensive and time-consuming. To address the shortcomings of these approaches, we propose a font generation method that employs the vanilla GAN approach to generate an infinite number of font styles but focuses on the real-time generation of photo-realistic font images. Additionally, we strive to create high-resolution images that can be used in practical applications. To accomplish this, we propose a conditional font GAN (CFGAN) with a sophisticated network architecture that is designed to generate novel style-consistent diverse font character sets. We control the generated characters in the proposed network using a non-trainable fixed character vector, while the style variation sampled from a Gaussian distribution is fused at all blocks of the generator through adaptive instance normalization (AdaIN) operation. Thus, the generator architecture can simultaneously generate an infinite number of font styles with style consistency and diversity during inference. We conducted various quantitative and qualitative experiments to demonstrate the effectiveness of the proposed model in terms of both image quality and computational cost.

Chinese Font Generation from Stroke Semantic and Attention Mechanism

In order to help designers develop computer Chinese fonts efficiently, a Chinese character font generation algorithm is proposed. The algorithm can generate full font character images with a relatively consistent style based on small font datasets. For the sake of solving problems in stroke adhesion and structural errors when using adversarial generative network (GAN), the prior information of stroke semantics is introduced to improve the deep neural network for alleviating wrong stroke generation. Moreover, a skip connection module equipped with attention mechanism is integrated in the network model to reduce incomplete font structure and training period. In this module, features in encoder can be projected to that in de- coder, which is helpful in eliminating the information loss of the decoder so as to avoid generating structural errors. Comparative experiments between the proposed algorithm and other methods including pix2pix, zi2zi and DCFont are conducted on public Chinese font datasets. The results show the proposed algorithm is better that the others from the perspective of SSIM and average in loss function. The industry application and scope of our algorithm are given in finally.

Integrating visual stimuli for enhancing neural text style transfer with EEG sensors

Font Designers need to create each font character by hand. With the help of what we propose, designers finish the process quickly and automatically. We use a neural network to learn and generate new fonts. We provide two vectors as input (Bigrams and Style Vectors) of Urdu language, encoded manually with one-hot encoding and t-distributed neighbor embedding. The transposed convolution neural network takes care of learning from input, where it decodes the input into beautiful fonts. Thus, by changing the style vector, the required changes are reflected in the resultant font style. Additionally, with the help of simulated annealing, we generate meaningful and full-length sentences. To evaluate whether the fonts generated are aesthetically sound, we provide the generated sentences to the end-users as visual stimuli and measure their responses in terms of their attention and meditation levels with EEG sensors. Higher sensor levels suggest the font quality and visual appeal.

Deep Neural Network Model-Assisted Reconstruction and Optimization of Chinese Characters in Product Packaging Graphic Patterns and Visual Styling Design

Chinese character fonts not only carry the long history of Chinese civilization, but also burst out modern design art elements with distinctive Chinese characteristics. This article first analyzes the origin and writing form of several ancient Chinese characters and draws out the influence of the historical evolution of ancient Chinese characters on Chinese culture. In the basic theoretical structure of font design, traditional art elements and modern font design are integrated, and specific design cases are analyzed. A Chinese character packaging quality detection method combining machine vision and a lightweight convolutional neural network is proposed. First, the method based on threshold segmentation and affine transformation in machine vision is used to perform threshold processing on the image to be tested, and the Chinese character region is tilted and cropped; then, the network structure of the classification algorithm is designed according to the requirements of image features and defect recognition. The field images are produced, a dataset of Chinese character packaging defects is established, and then the proposed Chinese character packaging defect recognition network is verified and deployed to test the accuracy and detection speed of the algorithm deployed on the Jetson Nano embedded platform. Combined with theoretical research and case analysis, the design of packaging design series is practiced with the idea of combining Chinese character art design and classical culture.

Large-Scale Printed Chinese Character Recognition for ID Cards Using Deep Learning and Few Samples Transfer Learning

In the field of computer vision, large-scale image classification tasks are both important and highly challenging. With the ongoing advances in deep learning and optical character recognition (OCR) technologies, neural networks designed to perform large-scale classification play an essential role in facilitating OCR systems. In this study, we developed an automatic OCR system designed to identify up to 13,070 large-scale printed Chinese characters by using deep learning neural networks and fine-tuning techniques. The proposed framework comprises four components, including training dataset synthesis and background simulation, image preprocessing and data augmentation, the process of training the model, and transfer learning. The training data synthesis procedure is composed of a character font generation step and a background simulation process. Three background models are proposed to simulate the factors of the background noise patterns on ID cards. To expand the diversity of the synthesized training dataset, rotation and zooming data augmentation are applied. A massive dataset comprising more than 19.6 million images was thus created to accommodate the variations in the input images and improve the learning capacity of the CNN model. Subsequently, we modified the GoogLeNet neural architecture by replacing the fully connected layer with a global average pooling layer to avoid overfitting caused by a massive amount of training data. Consequently, the number of model parameters was reduced. Finally, we employed the transfer learning technique to further refine the CNN model using a small number of real data samples. Experimental results show that the overall recognition performance of the proposed approach is significantly better than that of prior methods and thus demonstrate the effectiveness of proposed framework, which exhibited a recognition accuracy as high as 99.39% on the constructed real ID card dataset.

Self-supervised Chinese font generation based on square-block transformation

The generation of Chinese fonts has attracted much attention due to its wide range of applications such as artistic font generation, personalized font design, and calligraphy generation. The predominated Chinese font generation methods are mainly based on unpaired data and deep generative models such as generative adversarial networks. However, the existing deep methods for Chinese font generation commonly ignore the special structure information of Chinese characters, resulting in the lack of guidance during the feature extraction, and usually suffer from the mode collapse issue during the training procedure. Thus, the generation quality of Chinese characters needs to be improved.In order to solve this problem, this paper proposes a squared-block transformation-based self-supervised method to guide the model network to extract features with higher quality, and thus the proposed method significantly improves the performance on the generation of Chinese character fonts. It should be pointed out that the suggested squared-block geometric transformation does not require any modification of existing model networks and any additional human-labor cost. Thus, it can be adapted to many existing deep generative models for Chinese font generation. The effectiveness of the suggested self-supervised method is demonstrated by a series of experiments.The experiment results show that when equipped with the suggested squared-block transformation-based method, the popular CycleGAN-based deep model for Chinese font generation can significantly improve the quality of generated characters and stabilize the training procedure, while reducing the mode collapse.Moreover, when compared to other existing deep methods, the proposed method also outperforms them in terms of four evaluation metrics such as the content accuracy, FID, $L_1$ loss, and IOU, as well as the quality of generated characters.

A Prototypical Network-Based Approach for Low-Resource Font Typeface Feature Extraction and Utilization

This paper introduces a framework for retrieving low-resource font typeface databases by handwritten input. A new deep learning model structure based on metric learning is proposed to extract the features of a character typeface and predict the category of handwrittten input queries. Rather than using sufficient training data, we aim to utilize ancient character font typefaces with only one sample per category. Our research aims to achieve decent retrieval performances over more than 600 categories of handwritten characters automatically. We consider utilizing generic handcrafted features to train a model to help the voting classifier make the final prediction. The proposed method is implemented on the ‘Shirakawa font oracle bone script’ dataset as an isolated ancient-character-recognition system based on free ordering and connective strokes. We evaluate the proposed model on several standard character and symbol datasets. The experimental results showed that the proposed method provides good performance in extracting the features of symbols or characters’ font images necessary to perform further retrieval tasks. The demo system has been released, and it requires only one sample for each character to predict the user input. The extracted features have a better effect in finding the highest-ranked relevant item in retrieval tasks and can also be utilized in various technical frameworks for ancient character recognition and can be applied to educational application development.

SkelGAN: A Font Image Skeletonization Method

In this research, we study the problem of font image skeletonization using an end-to-end deep adversarial network, in contrast with the state-of-the-art methods that use mathematical algorithms. Several studies have been concerned with skeletonization, but a few have utilized deep learning. Further, no study has considered generative models based on deep neural networks for font character skeletonization, which are more delicate than natural objects. In this work, we take a step closer to producing realistic synthesized skeletons of font characters. We consider using an end-to-end deep adversarial network, SkelGAN, for font-image skeletonization, in contrast with the state-of-the-art methods that use mathematical algorithms. The proposed skeleton generator is proved superior to all well-known mathematical skeletonization methods in terms of character structure, including delicate strokes, serifs, and even special styles. Experimental results also demonstrate the dominance of our method against the state-of-the-art supervised image-to-image translation method in font character skeletonization task.

儿旁의 几旁과의 편방 혼동 현상과 관련 한자의 자형 표준화에 대한 提言

In Kaishu style, “人 rén” is transformed to “儿 rén” when it is located below another character. This varied form “儿 rén” is often written as “几 jī” when it is not attached to the part right above it. For this reason, the Chinese characters containing “儿 rén” that is not in contact with the upper part, such as “虎 hǔ”, “微 wēi”, “禿 tū”, “宂 rǒng”, “亮 liàng” and “兜 dōu”, often have a variant form in which “儿 rén” is transformed to “几 jī”.<BR> There are not a few inconsistencies of the forms of components in Chinese character fonts currently used in Korea; more seriously, even the same character has slightly different forms depending on the font. These inconsistencies have been an obstacle that hinders Chinese character education. The same problem is found in the characters that contain “儿 rén”. Therefore, this paper closely examined how neighboring countries in the Chinese character cultural sphere handled this issue for the establishment of standard Chinese character forms. In addition, this paper presents two suggestions that can serve as a reference in dealing with this issue in the process of revising domestic Chinese fonts based on the examples of other East Asian countries. First, based on Shuowenjiezi, for the characters containing “人 rén” in Xiaozhuan style, the forms with “儿 rén” are regarded as a standard in Kaishu style as well. Second, in consideration of the findings of Chinese grammatological studies and the tradition of Kaishu style in actual calligraphic practice, the forms with “几 jī” are regarded as a standard except for “宂 rǒng”, “呪 zhòu” and other characters that have been clearly known to include “儿 rén” for etymological reasons.<BR> The most important point is, whichever of the two is adopted, to improve the systemicity and regularity in the structure of Chinese characters by unifying the different component forms to a fullest extent. As the current issue, that the forms of the same component change slightly depending on the character or the forms of characters vary depending on the font, has a very negative effect on Chinese character education, an improvement is urgently needed.

Kansei Engineering Research for Design of Character Fonts and Comparison of Font Selection Methods

Kansei Engineering Research for Design of Character Fonts and Comparison of Font Selection Methods

Text Extraction and Recognition in Natural Scene Images using Contourlet Transform and PNN

Of late, the rapid development in the technology and multimedia capability in digital cameras and mobile devices has led to ever increasing number of images or multi-media data to the digital world. Particularly, in natural scene images, the text content provides explicit information to understand the semantics of images. Therefore, a system developed for extracting and recognizing texts accurately from natural scene images, in real-time, has significant relevance to numerous applications such as, assistive technology for people with vision impairment, tourist with language barrier, vehicle number plate detection, street signs, advertisement bill boards, robotics, etc. The extraction of the texts from natural scene images is a formidable task due to large variations in character fonts, styles, sizes, text orientations, presence of complex backgrounds and varying light conditions. The main focus of this research paper is to propose a novel hybrid approach for automatic detection, localization, extraction and recognition of text in natural scene images with cluttered background. Firstly, image regions with text are detected using edge features (GLCM) extracted from Contourlet transformed image and SVM (Support Vector Machine) classifier. Secondly, horizontal projection is applied on text regions for segmenting lines and vertical projection is applied on each text line for segmenting characters. The proposed method for text extraction has produced the precision, recall, F-Score and accuracy of 98.50%, 90.85.62%, 95.00%, and 89.90%, respectively. And, these results prove that the proposed method is efficient. Further, the so extracted characters are processed for recognition using contourlet transform and Probabilistic Neural Network (PNN) classifier. The computed features are moment invariants. Only the English script is considered for the experimentation. The proposed character recognition method has accuracy of 79.07%, which is higher in comparison to accuracy of 75.15% obtained by KNN (K-Nearest Neighbors) classifier

PUBLISHING SIGN OF UKRAINE OF THE SECOND HALF OF XX – BEGINNING OF XXI CENTURY: TRANSFORMATION IN TIME

Annotation. The article analyzes the little-studied art of the Ukrainian publishing sign, which has a history of almost fifty years and is well known for its highly artistic works. An excursion into the past of the Ukrainian publishing label has been made, collecting documents scattered across various sources of reports on samples of the publishing labels, both known and in the vast majority of unknown authors, and we pay tribute to this important cultural heritage that has developed in the course of the European process. The publishing signs of the second half of XX – beginning of XXI centuries are described and their artistic features are analyzed. The art of the Ukrainian emblem, especially in the last century, has not been practically studied. Probably, the topic is considered too small to draw enough attention. Turning pages of the book, few people pay atten- tion to the publishing house, thanks to which we have the happy opportunity to hold it. The artistic decision of modern publishing signs, which often quite often has a small font character, is not striking. They are created mainly by artists, editorial staff, without paying much attention to this process. However, there are times when they turn out to be the work of talented young artists who, over time, become famous. For example, the well-known art publishing signs for the leading Kiev publishing houses &amp;quot;Art&amp;quot; (artist V. E. Perevalsky) and &amp;quot;Rainbow&amp;quot; (artist O. I. Gubarev), created at that time by young graphic artists, who are now well-known folk artists of Ukraine. However, the art of the sign is an important component of book graphics, it does not lose its relevance at a new stage of society and deserves in-depth study.