Korean Font Research Articles

Robustness of Contrastive Learning on Multilingual Font Style Classification Using Various Contrastive Loss Functions

Font is a crucial design aspect, however, classifying fonts is challenging compared with that of other natural objects, as fonts differ from images. This paper presents the application of contrastive learning in font style classification. We conducted various experiments to demonstrate the robustness of contrastive image representation learning. First, we built a multilingual synthetic dataset for Chinese, English, and Korean fonts. Next, we trained the model using various contrastive loss functions, i.e., normalized temperature scaled cross-entropy loss, triplet loss, and supervised contrastive loss. We made some explicit changes to the approach of applying contrastive learning in the domain of font style classification by not applying any image augmentation. We compared the results with those of a fully supervised approach and achieved comparable results using contrastive learning with fewer annotated images and a smaller number of training epochs. In addition, we also evaluated the effect of applying different contrastive loss functions on training.

Detection Model of Hangul Stroke Elements: Expansion of Non-Structured Font and Influence Evaluation by Stroke Element Combinations

With the increase of various media, fonts continue to be newly developed. In Korea, numerous ‘Hangul’ fonts are also being developed, and as a result, the need for research on determining the similarity between fonts is emerging. For example, when creating a document, the font to be used must be downloaded from each computing environment. However, this is a very cumbersome process. If there is a font that is not supported in the system, the above problem can be easily solved by recommending the most similar font that can replace it. According to this need, we conducted various prior studies for similar font recommendations. As a result, we developed a ‘stroke element’ that exists in each consonant and vowel in Korean font and developed a font recommendation model using a stroke element. However, there is a limitation in that the existing research was studied only for the structured fonts corresponding to the printed type. Additionally, the font size was not considered in the font recommendation. In this study, two experiments were conducted to expand the font recommendation model by supplementing the limitations of existing studies. First, in order to enable similar font recommendations based on the stroke element even in fonts with various shapes, the font was classified according to the shape, and the stroke elements in each classification were detected. Second, when the font sizes were different, the change in the font recommendations result based on the stroke element was analyzed. In conclusion, we found that it was necessary to find a plan to extract stroke elements for font recommendation of fonts that do not belong to standard fonts. In addition, since the influence of the stroke element varies depending on the size of the font, we propose a stroke element weight model that can be used for recommendation by reflecting it.

글꼴의 네모틀 모양 및 삐침이 한글 낱자와 단어에 대한 시각 탐색에 미치는 영향

This study examined the influence of squared and serif fonts, which are representative visual properties of Korean fonts, on visual search for Korean letters and words. Korean letters and two-syllable words were used for search stimuli in each Experiment 1 and 2, and in both experiments, the distractors were manipulated to have the opposite font characteristic against the characteristic of a target while search accuracies and RTs were measured. The number of search items also varied to 4, 8, and 12 along which the search performance was measured. The results of both experiments showed it easier to search for a target without a squared font (among distractors with it) than the target with the squared font (among distractors without it), and to search for a target without serifs (among distractors with them) than the target with the serifs (among distractors without them), and these advantages were more evident in larger setsizes. The influence of font characteristics was also greater in Experiment 2 where two-syllable words were used for search stimuli than in Experiment 1 for a single Korean letter. These results indicate that searching a target without a square shape or serifs would be generally easier, and further suggest the possibility that the square and serif fonts can influence visual saliency of Korean letters and words.

SKFont: skeleton-driven Korean font generator with conditional deep adversarial networks

SKFont: skeleton-driven Korean font generator with conditional deep adversarial networks

Korean Font Synthesis with GANs

Korean Font Synthesis with GANs

Unbalanced U-Net과 GAN(Generative Adversarial Networks)을 이용한 한국어 폰트 자동 변환

본 논문에서는 원문 폰트를 특정한 아날로그 폰트 스타일로 변환하는 타이포그래피 변환 문제에 대해 연구한다. 타이포그래피 변환 문제를 해결하기 위해 이 문제를 이미지와 이미지 번역 문제로 치환하고 GAN을 기반으로 한 언밸런스 형 u-net 아키텍처를 제안한다. 기존의 밸런스 형 u-net과는 달리 제안하는 아키텍처는 언밸런스 형 u-net을 포함한 두 개의 서브넷으로 구성된다. (1)언밸런스 형 u-net은의미 및 구조 정보를 유지하면서 특정 글꼴 스타일을 다른 스타일로 변환한다. (2) GAN은 L1 손실, 상수손실 및 원하는 목표 글꼴을 생성하는 데 도움이 되는 이진 GAN 손실을 포함하는 복합 손실 함수를 사용한다. 실험결과 제안하는 모델인 언밸런스 형 u-net이 밸런스 형 u-net 보다 cheat loss에서 빠른 수렴속도와 안정적인 트레이닝 손실을 보였고 generate loss에서 트레이닝 손실을 안정적으로 줄여서 모델 성능 하락 문제를 해결하였다.

Algorithm of the Korean Character font design with 1458 Alphabet Combination Method

한글 음절 1만 1172개를 일정한 크기의 네모틀 안에서 정확하고 아름답게 구현하는데 728개 자소 조합 알고리즘과 903개 자소 조합 알고리즘이 사용되고 있으나 출판용 고품질 폰트를 디자인하려면 더 많은 자소의 조합 알고리즘이 필요하다. 따라서 한 개의 음절 공간 안에서 자소의 형태와 크기, 위치 등 공통 정보를 검토하여, 72개 그룹으로 구분한 다음 해당 자소와 대표 음절을 추출했다. 72그룹 알고리즘은 그룹별로 대표 자소를 추출하면 모음 자소 126개, 초성 자소 684개, 받침 자소 648개를 합친 1458개 자소가 나오므로 1458개 조합 알고리즘이라 부른다. 1458개 조합 알고리즘은 받침의 종류를 3가지로 대분하고, 모음을 12계열 알고리즘으로 구분한 다음 받침 종류에 따라 다시 36개의 소그룹으로 분류한 후 각기 초성의 종류(단자음/쌍자음)에 따라 2가지씩으로 세분하여 72개로 구분하였다. 한글 자소는 한 개의 음절 내에서 자소의 위치 정보에 따라 모양과 크기가 변한다.Usually 728 or 903 alphabet combination method is using for the Korean character font design. But designing more beautiful Korean character font for publishing books needs more than 903 alphabet combination. To analyze Korean alphabet syllables by classifying them into a new combination method by examining previous studies and approach Korean font design creatively and practically. According to the research purpose, the location information of 11,172 modern Korean character syllables was checked one by one and classified into 72 series and entered into the database file. 72 series classification divided initial sounds into singular consonants and geminate and further subdivided consonants placed under a vowel into no consonant placed under a vowel, singular consonants placed under a vowel, consonant clusters placed under a vowel. The following results were obtained through this study: first, in order to effectively design 11,172 Korean character syllables, 126 vowels, 684 initial sounds, 648 consonants placed under a vowel, a total of 1,458 phonemes are required according to 72 series classification. Korean character has a 3-dimensional characteristics that the location information of an initial sound, medial sound, consonant placed under a vowel is gathered to create one new syllable and hence, each location of the initial sound, medial sound, consonant placed under a vowel has a great influence on the remaining location information. Designing a Korean character syllable, it cannot be beautifully designed only with shape information, the aesthetic design element and location information, the intellectual design element is additionally required.

Analysis of font legibility for Korean serif and sans serif types in printed and displayed settings

Activities with smart devices have been more complicated and diversified, and font legibility influences these activities both directly and indirectly. This study investigated the difference of Korean font legibility between printed paper and VDT. Four Korean fonts, Dodum and Gulim - sans serif types, and Batang and Gungseo - serif types, were compared in experiments. The result showed that there was a significant difference in the font legibility for printed paper, but not for VDT. On printed paper, serif types were more legible than sans serif types. However, a significant difference between serif types and sans serif types were not found in VDT. Dodum, one of the sans serif font types achieved the highest score on preference in the survey questionnaire among the four Korean fonts. The paper suggests that serif types did not always provide the optimal legibility in VDT.

Effects of tinnitus and hearing loss on functional brain networks involved in auditory and visual short-term memory

Brain imaging data were acquired from three subject groups—persons with hearing loss and tinnitus (TIN), individuals with similar hearing loss without tinnitus (HL) and those with normal hearing without tinnitus (NH)—to test the hypothesis that TIN and control subjects use different functional brain networks for short-term memory. Previous studies have provided evidence of a link between hearing disorders such as tinnitus and the reorganization of auditory and extra-auditory functional networks. Greater knowledge of this reorganization could lead to the development of more effective therapies. Data analysis was conducted on fMRI data obtained while subjects performed short-term memory tasks with low or high attentional loads, using both auditory and visual stimuli in separate scanning sessions. Auditory stimuli were pure tones with frequencies between 500 and 1000 Hz. Visual stimuli were Korean fonts, unfamiliar to the subjects. We found similar behavioral response across the three groups for both modalities and tasks. However, the groups differed in their brain response, with these differences being more marked for the auditory tasks and not for the tasks involving visual stimuli.

Introducing Two Chinese/Japanese/Korean (CJK) Web Utilities

The author introduces websites created to enable users to read Chinese, Japanese, and Korean font.

What are Legible Korean Font Sizes within In-Vehicle Information Systems?

Objective: The aim of this study is to determine legible Korean font sizes within in-vehicle information systems(IVISs) in diving conditions. Background: Font legibility within IVISs is one of important causes on its" safe operations during driving. Several researches proposed some guidelines on the legible English font sizes within IVISs. On the contrary, appropriate Korean font sizes have been hardly known in spite of the typological differences between English and Korean. Therefore, more systematic researches for improving the legibility on Korean font size within IVISs have been required. Method: In this study, an experiment was performed with the following experimental factors: the existence of vibration, the color contrasts(white on black, black on white), the font types(HDR, CubeR, Gothic), and the font sizes(6, 8, 10, 12, 14, 16, 18, 20, 22, 24pt). To fit the experimental conditions into real driving environments, the illuminance was controlled to 15lx by using LED lamp and the distance between IVIS and participants was kept to 70cm. Moreover, all participants took the shutter glasses for employing well-known occlusion techniques. Results: The experimental results showed that "HDR" and "Non-vibration + Black on white" group took the shortest response time, and decreasing slopes of the response time with increasing font sizes were slowing down at 14pt then flattened out at 22pt regardless of the existence of vibration and color contrasts. Conclusion: The minimum size for legible Korean font would be about 14pt(5.47mm) and the optimum size would be about 22pt(8.59mm). Application: The guideline on the Korean font sizes from this study will be applied to design an IVIS in the future.

Fashion Graphic Design Based on National Brand of Korea - with a Focus on Pattern Design Utilizing Korean Font

Fashion Graphic Design Based on National Brand of Korea - with a Focus on Pattern Design Utilizing Korean Font

Aspects of the Development of Korean Font Design in the Digital Era

Aspects of the Development of Korean Font Design in the Digital Era

A Study on the analysis of Korean font classification System

A Study on the analysis of Korean font classification System