동물 캐릭터 적합 배우 추천 알고리즘 연구

이영숙 한국멀티미디어학회 2023 멀티미디어학회논문지 Vol.26 No.7

This study designed a virtual casting system that can be incorporated into the film industry. Through CNN deep learning and PCA region extraction we obtain face standard data for each animal and store similarity and region vector values in DB. We also leverage two similarity measurements to enable users to accurately provide the desired information. When the user chooses an animal of a character to be implemented in due diligence the recommender system utilizes the existing analyzed DB to provide standardized animal images. This allows users to recommend and cast actors who are most similar to the image of an implemented animal. This work is meaningful in showing that the results of existing studies can be utilized industrially.

청각장애인의 메타버스 콘텐츠 감상을 위한 보조 기능의 사용성 분석 -대중음악 콘서트를 중심으로-

김남희,임순범 한국멀티미디어학회 2023 멀티미디어학회논문지 Vol.26 No.8

Hearing-impaired people should be able to enjoy virtual reality content without discrimination. The development of metaverse has led to the success of virtual environment contents, including virtual concerts, but services for the deaf are still insufficient in this environment. In this paper, we present an auxiliary solution for hearing-impaired people to enjoy metaverse concerts and made video test samples for usability evaluation. To demonstrate that the solutions presented in this study help hearing-impaired people enjoy metaverse concerts, we made video test materials that provide secondary functions such as motion captions and haptic vibration. Test materials that provide basic subtitles were compared with test materials that applied motion captions, haptic vibrations, and both. To this end, an effectiveness evaluation and satisfaction evaluation using the Likert scale were conducted, and then a subjective satisfaction-oriented interview was conducted. Providing motion captions and haptic vibration together resulted in a relatively high usability evaluation effectiveness score, and specific requirements, problems, and improvements were identified through user interviews. If actual programs are developed and distributed based on this study, the quality of cultural life can be improved in the virtual environment of the hearing-impaired people.

딥러닝을 이용한 재무와 비재무 정보 기반 기업부도 예측 분석에 관한 연구

박중현 한국멀티미디어학회 2023 멀티미디어학회논문지 Vol.26 No.8

In the past, research related to corporate bankruptcy has primarily conducted empirical analyses through bankruptcy prediction models using financial ratios. However, with the advancement of ICT technology, there has been a growing trend in applying artificial intelligence. In this study, both traditional corporate bankruptcy prediction methodologies and machine learning and deep learning methodologies from the field of deep learning were applied to present the results of corporate bankruptcy prediction models and their predictive power. The dataset used included corporate characteristics, including financial ratios and non-financial information, as well as macroeconomic indicators to account for economic conditions. Five models, SVM, RF, DNN, CNN, and LSTM, were designated, and the model reliability and prediction accuracy for each model were analyzed. The LSTM model demonstrated superior performance and the highest prediction accuracy among the models. When comparing different approaches using only financial ratios (Set 1), using financial ratios and corporate characteristics together (Set 2), and incorporating financial ratios, corporate characteristics, and macroeconomic indicators (Set 3), which included all of these factors, consistently exhibited the highest model reliability and prediction accuracy.

다중 파라미터 MR 영상에서 유사 의료 영상으로 미세 조정한 자기지도학습 모델을 사용한 전립선암 악성도 예측 성능 개선

신예진,이민진,홍헬렌,황성일 한국멀티미디어학회 2023 멀티미디어학회논문지 Vol.26 No.8

In this paper, we propose a prostate cancer aggressiveness prediction model using self-supervised learning based on SimCLR with multi-parametric MR images. Self-supervised learning model is initially trained on the STL10 dataset, and then fine-tuned on the ProstateX dataset, which is similar to the downstream task dataset. To predict prostate cancer aggressiveness, downstream tasks are performed using each sequence of images from the multi-parametric MR dataset. The predicted results are combined using either majority voting or average voting for ensembling. Experimental results demonstrate that the self-supervised learning model fine-turned with similar images improves the performance by an average of 4.56% in accuracy, 20.69% in sensitivity, and 12.02% in negative predictive value. The ensemble method using majority voting with the self-supervised learning model fine-turned on similar images from the multi-parametric MR dataset yields the best performance in terms of accuracy (72.58%), balance accuracy (72.16%), and sensitivity (67.86%).

GAN을 이용한 웹툰 배경 이미지의 생성과 분석

이제경,김정기,안정인,임지연,차경애 한국멀티미디어학회 2023 멀티미디어학회논문지 Vol.26 No.8

In this paper, we propose a method to reduce the amount of manual work in webtoon creation and utilize creative contents derived from AI learning through deep learning-based technology that generates background images of various styles. To achieve this goal, we train CartoonGAN and AnimeGAN models that are specialized in creating images in the style of webtoons and animations, and create background images that can be used for webtoons. Recently, various Generative Adversarial Network (GAN) models have been actively used to create digital content, but cartoon-style images should be created with simplified textures and sharp outlines. In addition, when converting a real image into a cartoon style, it is necessary to create a simple and abstract image while maintaining the content expressed by the image. We build training data suitable for the production of these webtoon-style images, and analyze whether the images generated by the two GAN models can be used for webtoon production, and seek ways to utilize generative AI.

다국적 각색 영화가 관객 만족도에 미치는 영향 연구 -<허삼관매혈기>를 각색한 영화 <허삼관>을 중심으로-

유원원,이연우,김치용 한국멀티미디어학회 2023 멀티미디어학회논문지 Vol.26 No.8

The Korean film <Heosamgwan> adapted from Yu Hua's novel <Chronicle of a Blood Merchant> is well received. Chinese and Korean scholars have made a lot of in-depth research on the adaptation of <Heosamgwan>, from the narrative text to the adaptation strategy, but few of them have focused on the audience. In this study, a survey on satisfaction was conducted after watching the movie <Heosamgwan> for Chinese international students in Korea, and a SEM model was built based on the satisfaction theory. The relationship between storytelling, audiovisual effects, and emotional experience was analyzed through the model.

결측치 보간 알고리즘을 적용한 앙상블 기반의 태양광 발전량 예측 시스템

박수빈,김진성,정세훈,심춘보 한국멀티미디어학회 2023 멀티미디어학회논문지 Vol.26 No.8

Environmental problems such as global warming due to excessive use of fossil fuels are becoming serious. In order to solve this problem, the supply of new and renewable energy is being activated, and the new and renewable energy market is also expanding. In particular, the share of solar and wind energy among new and renewable energies is rapidly increasing. However, uncertainty and volatility are inherent in renewable energy due to the characteristics of power generation that depend on natural conditions. This leads to a problem in which errors occur in the prediction of the amount of reserve energy required to secure the amount and cost of renewable energy generation. In this paper, we propose an ensemble-based solar power generation prediction system applying missing value interpolation algorithm. It predicts the amount of solar power generation by using weather forecast data from the Korea Meteorological Administration, and provides visualization and scheduling functions for the amount of power generation and predicted amount through a web page.

국내 모바일 성범죄 예방 어플리케이션 사용성 분석

최지영,권만우,김치용,황미경 한국멀티미디어학회 2023 멀티미디어학회논문지 Vol.26 No.7

Sexual crimes account for a significant proportion of total crimes and originate from various types of offenses. Although smartphone-based applications for preventing sexual crimes have been developed, they have not been extensively analyzed or promoted, resulting in low public awareness. In this study, we analyzed the strengths, weaknesses, and features of the top five applications based on App Store ratings as of January 2023, using reconstructed analysis indicators to identify application components. The components of the analysis indicators were categorized into functionality, informativeness, and design. Functionality included reporting features and user protection functions, informativeness encompassed location information and offender information, and design involved UI and UX. The results revealed that the applications offered various reporting methods in the functional area and useful features in the informative area, such as providing offender information and real-time safe route guidance based on the user's location. However, they were limited due to their availability in specific regions only. In the design area, the applications were designed to align with their characteristics but presented some usability issues. The findings of this study are expected to serve as a reference for the development and promotion of domestic mobile applications for the prevention of sexual crimes based on smartphones.

시점 변경을 위한 깊이 정보를 이용한 홀채움 기법

이예솝 한국멀티미디어학회 2023 멀티미디어학회논문지 Vol.26 No.7

The significance of precise image inpainting for image view synthesis has grown in recent years due to the rapid development of autonomous driving or XR (eXtended Reality) devices or robotics. Numerous studies have been conducted on image inpainting during this time, but they have involved intricate procedures and neglected depth information. This paper introduces a novel and straightforward hole-filling algorithm that leverages local depth information. We generate direction vector map based on the local depth information for each hole pixel and utilize the map for the image inpainting. We perform experiments using two distinct simple inpainting methods and compare them with the method augmented by depth information, evaluating both qualitative and quantitative metrics. The results of these experiments demonstrate improvements in numerical accuracy and plausibility.

사용자 참여를 위한 어포던스 요소를 기반으로 혼합현실 지역 문화 실감형 콘텐츠 개발

최하람,남상훈 한국멀티미디어학회 2023 멀티미디어학회논문지 Vol.26 No.7

Realistic content based on immersive technology stimulates the human senses and enables a realistic experience like reality. Realistic content can increase immersion and presence, which can induce users to participate in the content. Accordingly, cultural institutions and museums are developing realistic content to enable users' sensory experience and provide immersion. Using Mixed Reality in art galleries and museums can provide users with additional information about the content. In addition, users can virtually experience artifacts and historical people that cannot be experienced at present. Affordance, which is important in user experience, can provide a high sense of immersion and interaction to users in realistic content. Therefore, it is necessary to consider affordance in developing realistic content. In this paper, regional-culture realistic content using mixed reality was developed. Sculptor Moonshin, who cannot be seen currently, was developed as virtual content. To participate user in the content, the content was developed based on Hartson's affordance, and the developed content was demonstrated and experienced in cooperation with Changwon City Moonshin Art Museum.