In this paper, we propose a semi-supervised domain adaptation solution to deal with practical face recognition (FR) scenarios where a single face image for each target identity (to be recognized) is only available in the training phase. Main goal of the proposed method is to reduce the discrepancy between the target and the source domain face images, which ultimately improves FR performances. The proposed method is based on the Domain Adatation network (DAN) using an MMD loss function to reduce the discrepancy between domains. In order to train more effectively, we develop a novel loss function learning strategy in which MMD loss and cross-entropy loss functions are adopted by using different weights according to the progress of each epoch during the learning. The proposed weight adoptation focuses on the training of the source domain in the initial learning phase to learn facial feature information such as eyes, nose, and mouth. After the initial learning is completed, the resulting feature information is used to training a deep network using the target domain images. To evaluate the effectiveness of the proposed method, FR performances were evaluated with pretrained model trained only with CASIA-webface (source images) and fine-tuned model trained only with FERET's gallery (target images) under the same FR scenarios. The experimental results showed that the proposed semi-supervised domain adaptation can be improved by 24.78% compared to the pre-trained model and 28.42% compared to the fine-tuned model. In addition, the proposed method outperformed other state-of-the-arts domain adaptation approaches by 9.41%.

As consumption of high-quality fruits increases and sales and packaging units become smaller, the demand for automatic fruit grading systems is increasing. Compared to other crops, the quality of fruit is determined by visual characteristics such as shape, color, and scratches, rather than just physical size and weight. Accordingly, this study presents a CNN model that can effectively extract and classify the visual features of fruits and a perceptron that classifies fruits using physical features, and proposes a stacking ensemble model that can effectively combine the classification results of these two neural networks. The experiments with AI Hub public data show that the stacking ensemble model is effective for grading fruits. However, the ensemble model does not always improve the performance of classifying all the fruit grading. So, it is necessary to adapt the model according to the kind of fruit.

Lately, over 4K high definition and high dynamic range (HDR) display devices are popularized, various interpolation and HDR methods have been researched to expand the size and the dynamic range. Since most of the legacy low resolution (LR) images require both an interpolation and a HDR tone mapping methods, the two processes should be subsequently applied. Therefore, the proposed algorithm presents a HDR up-scaling algorithm using undecimated wavelet transform and Retinex method, which transfers a LR image of low dynamic range (LDR) into the high resolution (HR) with HDR. The proposed algorithm consists of an up-scaling scheme increasing the image size and a tone mapping scheme expanding the dynamic range. The up-scaling scheme uses the undecimated version of the simplest Haar wavelet analysis for the 8-directional interpolation and the change region is extracted during the analysis. This region information is utilized in controlling the surround functions' size of the proposed tone mapping using MSRCR, to enhance the pixels of around the edges that are dominant feature of the subjective image quality. As the results, the proposed algorithm can apply an up-scaling and tone mapping processes in accordance with the type of pixel.

The 2-dimensional arrangement method of nodes has been used in most of RF (Radio Frequency) based communication network simulations. However, this method is not useful for the an none-obstacle 3-dimensional space networks in which the propagation delay speed in communication is very slow and, moreover, the values of performance factors such as the communication speed and the error rate change on the depth of node. Such a typical example is an underwater communication network. The 2-dimensional arrangement method is also not useful for the RF based network like some WSNs (Wireless Sensor Networks), IBSs (Intelligent Building Systems), or smart homes, in which the distance between nodes is short or some of nodes can be arranged overlapping with their different heights in similar planar location. In such cases, the 2-dimensional network simulation results are highly inaccurate and unbelievable so that they lead to user's erroneous predictions and judgments. For these reasons, in this paper, we propose a method to place uniformly and randomly communication nodes in 3-dimensional network space, making the wireless link with neighbor node possible. In this method, based on the communication rage of the node, blocks are generated to construct the 3-dimensional network and a node per one block is generated and placed within a block area. In this paper, we also introduce an algorithm based on this method and we show the performance results and evaluations on the average time in a node generation and arrangement, and the arrangement time and scatter-plotted visualization time of all nodes according to the number of them. In addition, comparison with previous studies is conducted. As a result of evaluating the performance of the algorithm, it was found that the processing time of the algorithm was proportional to the number of nodes to be created, and the average generation time of one node was between 0.238 and 0.28 us. ultimately, There is no problem even if a simulation network with a large number of nodes is created, so it can be sufficiently introduced at the time of simulation.

The access control system is a system that allows users to selectively enter the building by granting an access key to the user for security. Access keys with weak security are easily exposed to attackers and cannot properly perform the role that authenticates users. Access code keys should be protected from forgery or spoofing. For this reason, access key verification service models is important in security. However, most models manage all access keys on one central server. This method not only interrupts all services due to server errors, but also risks forgery and spoofing in the process of transmitting access keys. Therefore, blockchain algorithms are used to reduce this risk. This paper proposes a blockchain-based access key verification service model that used distributed stored blockchain gateways on storing access keys and authenticates the user's identity based on them. To evaluate the performance of this model, an experiment was conducted to confirm the performance of the access key forgery recovery rate and the blockchain network performance. As a result, the proposed method is 100% forgery recovery rate, and the registration and verification process is evaluated at 387.58 TPS and 136.66 TPS.

Recently, as the demand for physical security technology to prevent leakage of technical and business information of companies and public institutions increases, the high tech companies are operating X-ray security checkpoints at building entrances to protect their intellectual property and technology. X-ray security checkpoints are operated to detect cameras and storage media that may store or leak important technologies in the bags of people entering and leaving the building. In this study, we propose an X-ray security checkpoint system that automatically detects a storage medium in an X-ray image using a deep learning based object detection method. The proposed system consists of an edge computing unit and a cloud-computing unit. We employ the RetinaNet for automatic storage media detection in the X-ray security checkpoint images. The proposed approach achieved mAP of 95.92% on private dataset.

This paper studies a model to diagnose efficiency reduction of inverter using Multilayer Perceptron(MLP). In this study, two inverter data which started operation at different day was used. A Multilayer Perceptron model was made to predict photovoltaic power data of the latest inverter. As a result of the model's performance test, the Mean Absolute Percentage Error(MAPE) was 4.1034. The verified model was applied to one-year-old and two-year-old data after old inverter starting operation. The predictive power of one-year-old inverter was larger than the observed power by 724.9243 on average. And two-year-old inverter's predictive value was larger than the observed power by 836.4616 on average. The prediction error of two-year-old inverter rose 111.5572 on a year. This error is 0.4% of the total capacity. It was proved that the error is meaningful difference by t-test. The error is predicted value minus actual value. Which means that PV system actually generated less than prediction. Therefore, increasing error is decreasing conversion efficiency of inverter. Finally, conversion efficiency of the inverter decreased by 0.4% over a year using this model.

The relationship between film and culture is established through a shared focus on people's lives. Culture changes at different times, which makes a dynamic and complex relationship between film and culture. Understanding this relationship helps to provide better support for current filmmaking. The purpose of this study is to provide insights into the macro issue of film and culture by exploring the relationship between film theme and culture concept. To this end, we used a case study to explore the connection between themes and culture concepts in Disney's Mulan (1998, 2020) through literature and analyses of culture concepts (Chinese "imagery" and Postmodernism). The study concludes that, "Culture assemblage" and "Vision is king" have become cultural trends that influence the creation of film theme and style to a certain extent; how to balance the relationship between the core idea of film theme and the external forms is an important issue in current filmmaking. The article contributes to promote thinking about culture in film creation, and its research perspective on different cultures and histories is innovative.

This is an empirical study to find out whether subtitles and infographics can affect the recognition and recall memory of YouTube users. In this study, a total of 104 university students were divided into 4 groups and exposed to 4 types of contents according to the presence or absence of subtitles and infographics. After the subjects watched YouTube content, they responded to the level of their memory through a survey. As a result of the analysis, the subtitles used in YouTube contents did not affect the recall memory of the audience, and the infographic did not affect the recognition memory of the audience. However, when subtitles were used for YouTube content, the audience's recognition memory was found to be high at a statistically significant level, and when infographics were used for YouTube content, the audience's recall memory was found to be high at a statistically significant level. The significance of this study can be found in that the effects of subtitles and infographics that appeared in the audience effect of legacy media such as newspapers and broadcasting were also found in YouTube, a new video content media.

Mickey Mousing perfect match between animation sound and image was an aesthetic in the field of animation, but since the 2000s, works such as and released by producers such as DreamWorks and Pixar have expanded the perfection of synchronization to irony. It also influenced the identification system of sentiment. It is time to view the directing attempt of these elements as a factor that changed the new paradigm of narrative, and related research is needed. In this study, the scene of was analyzed as a case study for the synchronization of animation sound and image components and the boundary direction on the recognition of identification between reality and fiction. Aesthetic recognition of the research work is based on the premise of real time and space perception, and the audience can recognize in the conceptual world as an integrated art by playfully producing fictional time and space. The direct antithesis of synchronization and identification was drawn to maintain the curiosity of the next scene by repeating selective concealment and disclosure of information in the direction of conveying an unfamiliar and heterogeneous feeling to the audience.

While public interests in the metaverse are growing recently in the Korean media and research, its understanding has not been fully established yet. In this study, we aimed to probe whether the rapid growth in media attention about the metaverse has increased its usage as a buzzword accompanied by an absence of scientific context. We analyzed publications and online news containing "metaverse" from 2020 to 2022. The data analysis methods are 1) time series frequency, 2) keyword network, 3) natural language model. The findings indicate the perception gap about metaverse between research and news articles broadened as its popularity has grown. Research about metaverse gradually expanded its connections with related topics-virtual and augmented realities-focusing on social changes in a remote environment. However, media reporting frequently used "metaverse" as a buzzword rather than explaining its scientific background, stimulating the proliferation of related topics and the dispersion of news content. This study further discusses the need for a media strategy to improve public conception of the long-term development of the metaverse.

In the context of globalization, circulation between cities has become more frequent. The airport is no longer just a place for boarding, disembarking, and transportation, but a public place that serves as the communication function of the "aviation city". The intervention of public art in the airport space not only gives users a sense of space experience, but also becomes a unique carrier for city and country image shaping. The purpose of this paper is to study the emotional value brought by airport public art to users, and to investigate the correlation analysis of public art design elements and user preferences based on this premise. The research methods are machine learning method and SPSS 21.0. The user's emotional value is introduced in the big data evaluation, and the preference and inclination of airport users to various elements of public art are analyzed by questionnaire. Through the research conclusion, the preference and main contradiction of users in the airport for the four dimensions of public art design elements are obtained. Opinions and optimization methods to provide reference data and theoretical support for public art design.

Artificial intelligence application in digital health care has been increasing with its development of artificial intelligence. The convergence of the healthcare industry and information and communication technology makes the diagnosis of diseases more simple and comprehensible. From the perspective of medical services, its practice as an initial test and a reference indicator may become widely applicable. Therefore, analyzing the factors that are the basis for existing diagnosis protocols also helps suggest directions using artificial intelligence beyond previous regression and statistical analyses. This paper conducts essential diagnostic prediction learning based on the analysis of blood cancer factors reported previously. Blood cancer diagnosis predictions based on artificial intelligence contribute to successfully achieve more than 90% accuracy and validation of blood cancer factors as an alternative auxiliary approach.

This study proposes the classification model of Minhwa genre based on object detection of deep learning. To detect unique Korean traditional objects in Minhwa, we construct custom datasets by labeling images using object keywords in Minhwa DB. We train YOLOv5 models with custom datasets, and classify images using predicted object labels result, the output of model training. The algorithm consists of two classification steps: 1) according to the painting technique and 2) genre of Minhwa. Through classifying paintings using this algorithm on the Internet, it is expected that the correct information of Minhwa can be built and provided to users forward.

Sign language has completely different meaning depending on the direction of the hand or the change of facial expression even with the same gesture. In this respect, it is crucial to capture the spatial-temporal structure information of each movement. However, sign language translation studies based on Sign2Gloss2Text only convey comprehensive spatial-temporal information about the entire sign language movement. Consequently, detailed information (facial expression, gestures, and etc.) of each movement that is important for sign language translation is not emphasized. Accordingly, in this paper, we propose Spatial-temporal Keypoints Centered Sign2Gloss2Text Translation, named STKC-Sign2 Gloss2Text, to supplement the sequential and semantic information of keypoints which are the core of recognizing and translating sign language. STKC-Sign2Gloss2Text consists of two steps, Spatial Keypoints Embedding, which extracts 121 major keypoints from each image, and Temporal Keypoints Embedding, which emphasizes sequential information using Bi-GRU for extracted keypoints of sign language. The proposed model outperformed all Bilingual Evaluation Understudy(BLEU) scores in Development(DEV) and Testing(TEST) than Sign2Gloss2Text as the baseline, and in particular, it proved the effectiveness of the proposed methodology by achieving 23.19, an improvement of 1.87 based on TEST BLEU-4.