• 한국정보과학회논문지:소프트웨어및응용
• /
• 제31권6호
• /
• pp.808-816
• /
• 2004

본 논문에서는 다중 특징 결합과 유사도 공간을 이용한 실제적인 온라인 얼굴 검증 시스템을 구현하는 방법을 제안한다. 얼굴 검증에서의 주요 쟁점은 다양한 얼굴 형상 변화의 처리이다. 이러한 변화는 단지 한가지 특징만으로는 해결되기 어렵다. 따라서 얼굴 형상에 있어서의 다양한 변화를 처리하기 위해서 상호보완적인 특징들의 결합이 필요하다. 이러한 관점에서 우리는 다중 주성분 분석과 에지 분포에 기반 한 특징 추출 방법을 제안한다. 이러한 특징들은 다수의 간단한 유사도 측정 방법들로 형성된 새로운 intra-person/extra-person 유사도 공간으로 사상되고, 최종적으로 Support Vector Machine에 의해 평가된다. 실제적인 대용량 데이터 베이스로 실험한 결과, equal error rate 0.029의 결과를 나타내었고, 이는 많은 실제 응용제품에도 충분히 팩용 가능한 수준이다.

• 한국멀티미디어학회논문지
• /
• 제24권8호
• /
• pp.1114-1121
• /
• 2021

In this paper, we implement a recipe recommendation system using ingredient harmonization analysis based on recipe data. The proposed system receives an image of a food ingredient purchase receipt to recommend ingredients and recipes to the user. Moreover, it performs preprocessing of the receipt images and text extraction using the OCR algorithm. The proposed system can recommend recipes based on the combined data of ingredients. It collects recipe data to calculate the combination for each food ingredient and extracts the food ingredients of the collected recipe as training data. And then, it acquires vector data by learning with a natural language processing algorithm. Moreover, it can recommend recipes based on ingredients with high similarity. Also, the proposed system can recommend recipes using replaceable ingredients to improve the accuracy of the result through preprocessing and postprocessing. For our evaluation, we created a random input dataset to evaluate the proposed recipe recommendation system's performance and calculated the accuracy for each algorithm. As a result of performance evaluation, the accuracy of the Word2Vec algorithm was the highest.

• International Journal of Computer Science & Network Security
• /
• 제24권6호
• /
• pp.1-7
• /
• 2024

Comparing text features involves evaluating the "similarity" between texts. It is crucial to use appropriate similarity measures when comparing similarities. This study utilized various techniques to assess the similarities between newspaper articles, including deep learning and a previously proposed method: a combination of Pointwise Mutual Information (PMI) and Word Pair Matching (WPM), denoted as PMI+WPM. For performance comparison, law data from medical research in Japan were utilized as validation data in evaluating the PMI+WPM method. The distribution of similarities in text data varies depending on the evaluation technique and genre, as revealed by the comparative analysis. For newspaper data, non-deep learning methods demonstrated better similarity evaluation accuracy than deep learning methods. Additionally, evaluating similarities in law data is more challenging than in newspaper articles. Despite deep learning being the prevalent method for evaluating textual similarities, this study demonstrates that non-deep learning methods can be effective regarding Japanese-based texts.

• 한국멀티미디어학회논문지
• /
• 제25권8호
• /
• pp.1059-1068
• /
• 2022

Recently, research using image-based deep learning is being conducted to determine similar fonts or recommend fonts. In order to increase the accuracy in judging the similarity of Hangul fonts, a previous study was conducted to calculate the similarity according to the combination of stroke elements. In this study, we tried to solve this problem by designing an integrated model that reflects the weights for each stroke element. By comparing the results of the user's font similarity calculation conducted in the previous study and the weighted model, it was confirmed that there was no difference in the ranking of the influence of the stroke elements. However, as a result of comparison by letter sizes, it was confirmed that there was a difference in the ranking of the influence of stroke elements. Accordingly, we proposed a weighted model set separately for each font size.

• 패션비즈니스
• /
• 제22권1호
• /
• pp.135-147
• /
• 2018

Since the 20th century, there has been a growing interest in the new concept of fractals, a combination of mathematics and art, and the attempt to study the creative spatial aspects of the concept is being made. The purpose of this research is to examine artistic characteristics of fractal dimension and then analyze the types of fractal dimensions expressed in the fashion. Previous literature on fractals and dimension, and visual data on art and fashion collected over the Internet were used for analysis. Fractal dimension refers to the spatial concept of structural dimension of geometrical self-similarity. An analysis of the types of fractals seen in fashion revealed spatial expansion, the repetition in continual figures, superposition accordant to different sizes, and shades of different shapes. The aesthetic characteristics of fractal dimension appearing in fashions were examined based on analyses of fractal dimension types; the inherent characteristics of self-similarity, superimposition, and atypicality were found. Results obtained from this study are expected to be used as basic materials for the application of the design of fractal dimension into various perspectives of fashion.

• 패션비즈니스
• /
• 제13권6호
• /
• pp.125-136
• /
• 2009

The main objective of this research was to understand the characteristics of complementary color scheme through the analysis of contemporary women's fashion color coordination as they appear in the 'Collections'. Data collection of 115 was done through review of 'pr$\hat{e}$t-$\grave{a}$-porter Collections' of four cities; Milan, London, New York, Paris. Statistical analysis of frequency and also qualitative interpretation of characteristics of complementary color harmony characteristics which appeared in four collections were completed. The main findings were as followed; (1) Only 115 complementary color schemes out of 4968 two color combinations appeared. Compared the comparative ratios of appearance frequency between the two color combination and the complementary color harmony, London showed the highest, and followed by New York, and Milan and Paris. (2)The combined color type of 'red+green' was the most frequently appeared, and followed by 'violet+yellow' and 'orange+blue'. For the type of tone harmony, the contrast tone showed the most, and followed by the similarity and identity. According to the type of complementary color combination, the type of tone harmony used differently. Some differences showed in the types of color harmony and tone harmony between collections. (3) The complementary color scheme which shown in four collections harmonized two opposite colors mainly through the strong tone contrast and this could cause tension along with interest in the image. The complementary colors intensified and brought out the attributes each other. More details, high contrast of two complementary colors of yellow and violet created a vibrant look especially when used at higher saturation. Sometimes, however, some tone variations of two hues neutralized the strong effect and sometimes enhanced each other. When they used in similarity tones or identity tones in light colors, the tension was reduced and became softened but still presented nice harmony. In the type of 'red+green' color harmony, the various color combinations were demonstrated, mostly through tone manipulation of green color. The similarity tone harmony, which used the most, could effect a better sense of harmony and present more sophisticated looks. When used in contrast tone harmony, some changes in its own color which have only one color of two the excessive intensity led a good harmony. The 'orange+blue' color harmony was shown the least and used three tone harmony almost the same ratio. In this color harmony, blue amplified its energy and brilliance of orange and seemed to work better when one color was at a lower intensity than the other. In harmony with a similarity and an identity tone, this color harmony produces a stable and calm image. (4) The complementary scheme appeared more frequently in the S/S collections than in A/W collections and showed some differences in the types of color harmony and tone harmony between seasons, however, no big differences between collections.

• 대한의용생체공학회:의공학회지
• /
• 제26권4호
• /
• pp.231-241
• /
• 2005

New methods to register multiple hemispheric slices of the postmortem brain to anatomically corresponding in-vivo MRI slices within a 3D volumetric MRI are presented. Gel-embedding and fiducial markers are used to reduce geometrical distortions in the postmortem brain volume. The registration algorithm relies on a recursive extraction of warped MRI slices from the reference MRI volume using a modified non-linear polynomial transformation until matching slices are found. Eight different voxel similarity measures are tested to get the best co-registration cost and the results show that combination of two different similarity measures shows the best performance. After validating the implementation and approach through simulation studies, the presented methods are applied to real data. The results demonstrate the feasibility and practicability of the presented co­registration methods, thus providing a means of MR signal analysis and histological examination of tissue lesions via co­registered images of postmortem brain slices and their corresponding MRI sections. With this approach, it is possible to investigate the pathology of a disease through both routinely acquired MRls and postmortem brain slices, thus improving the understanding of the pathological substrates and their progression.

• 환경영향평가
• /
• 제18권6호
• /
• pp.401-409
• /
• 2009

Multivariate statistical approaches to classify sampling sites with measuring their similarity by water quality data and understand the characteristics of classified clusters have been discussed for the optimal water quality monitering network. For empirical study, data of two years (2005, 2006) at the 9 sampling sites with the combination of 2 depth levels and 7 important variables related to water quality is collected in Yongdam reservoir. The similarity among sampling sites is measured with Euclidean distances of water quality related variables and they are classified by hierarchical clustering method. The clustered sites are discussed with principal component variables in the view of the geographical characteristics of them and reducing the number of measuring sites. Nine sampling sites are clustered as follows; One cluster of 5, 6, and 7 sampling sites shows the characteristic of low water depth and main stream of water. The sites of 2 and 4 are clustered into the same group by characteristics of hydraulics which come from that of main stream. But their changing pattern of water quality looks like different since the site of 2 is near to dam. The sampling sites of 3, 8, and 9 are individually positioned due to the different tributary.

• 한국수자원학회:학술대회논문집
• /
• 한국수자원학회 2010년도 학술발표회
• /
• pp.957-961
• /
• 2010

Multivariate statistical approaches to classify sampling sites with measuring their similarity by water quality data. For empirical study, data of two years at the 9 sampling sites with the combination of 2 depth levels and 7 important variables related to water quality is collected in reservoir. The similarity among sampling sites is measured with Euclidean distances of water quality related variables and they are classified by hierarchical clustering method. The clustered sites are discussed with principal component variables in the view of the geographical characteristics of them and reducing the number of measuring sites. Nine sampling sites are clustered as follows; One cluster of 5, 6, and 7 sampling sites shows the characteristic of low water depth and main stream of water. The sites of 2 and 4 are clustered into the same group by characteristics of hydraulics which come from that of main stream. But their changing pattern of water quality looks like different since the site of 2 is near to dam. The sampling sites of 3, 8, and 9 are individually positioned due to the different tributary.

• Journal of Photoscience
• /
• 제7권2호
• /
• pp.73-78
• /
• 2000

To analyze molecular traits determining pigmentation between Pharbitis nill violet and white, Amplified Fragment Length Polymorphism(AFLP) and Differential Display Reverse Transcription(DDRT) experiments were carried out with either genomic DNAs or total RNAs isolated from both plants. Results of AFLP experiment in combination of 8 EcoRⅠ primers with 6 MseⅠ primers showed 41 violet-and 60 white-specific DNA bands. In the subsequent experiment, 22 violet-and 22 white-specific DNA fragments were amplified by PCR with DNAs eluted. The sizes of the fragments range from 200 to 600bp. DDRT using total RNA produced 19 violet-and 17 white-specific cDNA fragments, ranging from 200 to 600bp. The fragments obtained by both AFLP and DDRT had been cloned into pGEM T-easy vector, amplified and subjected to the nucleotide sequence analyses. As a result of Blast sequence analysis, most of them sequenced up to date showed no similarity to any Known gene, while few has similarity to known animal or plant genes. An AFLP clone V6, for example, has a strong sequence similarity to the human transcription factor LZIP-alpha mRNA and a DDRT clone W19 to Solanum tuberosum 3-hydroxy-3-methylglutaryl coenzyme A reductase mRNA.