• The Journal of Society for e-Business Studies
• /
• v.26 no.1
• /
• pp.29-41
• /
• 2021

Recently, there are increasing attempts to utilize deep learning methodology in the fashion industry. Accordingly, research dealing with various fashion-related problems have been proposed, and superior performances have been achieved. However, the studies for fashion style classification have not reflected the characteristics of the fashion style that one outfit can include multiple styles simultaneously. Therefore, we aim to solve the multi-label classification problem by utilizing the dependencies between the styles. A multi-label recognition model based on a graph convolution network is applied to detect and explore fashion styles' dependencies. Furthermore, we accelerate model training and improve the model's performance through transfer learning. The proposed model was verified by a dataset collected from social network services and outperformed baselines.

• Journal of the Korean Society of Fisheries and Ocean Technology
• /
• v.57 no.2
• /
• pp.149-161
• /
• 2021

The concern on the greenhouse gas emission is strongly increasing globally. In fishery industry section, the greenhouse gas emissions are an important issue according to The Paris Climate Change Accord in 2015. The Korean government has a plan to reduce the GHG emissions as 4.8% compared to the BAU in fisheries until 2020. Furthermore, the Korean government has also declared to achieve the carbon neutrality in 2050 at the Climate Adaptation Summit 2021. However, the investigation on the GHG emissions from Korean fisheries did not carry out extensively. Most studies on GHG emissions from Korean fishery have dealt with the GHG emissions by fishery classification so far. However, follow-up studies related to GHG emissions from fisheries need to evaluate the GHG emission level by species to prepare the adoption of environmental labels and declarations (ISO 14020). The purpose of this research is to investigate which degree of GHG emitted to produce the species (hairtail and small yellow croaker) from various fisheries. Here, we calculated the GHG emission to produce the species from the fisheries using the Life Cycle Assessment method. The system boundary and input parameters for each process level are defined for the LCA analysis. The fuel use coefficients of the fisheries for the species are also calculated according to the fuel type. The GHG emissions from sea activities by the fisheries will be dealt with. Furthermore, the GHG emissions for producing the unit weight species and annual production are calculated by fishery classification. The results will be helpful to understand the circumstances of GHG emissions from Korean fisheries.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.41 no.3
• /
• pp.277-288
• /
• 2021

Building information modeling (BIM) element to industry foundation classes (IFC) entity mappings need to be checked to ensure the semantic integrity of BIM models. Existing studies have demonstrated that machine learning algorithms trained on geometric features are able to classify BIM elements, thereby enabling the checking of these mappings. However, reliance on geometry is limited, especially for elements with similar geometric features. This study investigated the employment of relational data between elements, with the assumption that such additions provide higher classification performance. Neural structured learning, a novel approach for combining structured graph data as features to machine learning input, was used to realize the experiment. Results demonstrated that a significant improvement was attained when trained and tested on eight BIM element types with their relational semantics explicitly represented.

• Journal of the Institute of Convergence Signal Processing
• /
• v.21 no.3
• /
• pp.121-126
• /
• 2020

Sound-based machine fault diagnosis is the automatic detection of abnormal sound in the acoustic emission signals of the machines. Conventional methods of using mathematical models were difficult to diagnose machine failure due to the complexity of the industry machinery system and the existence of nonlinear factors such as noises. Therefore, we want to solve the problem of machine fault diagnosis as a deep learning-based image classification problem. In the paper, we propose a CNN-based automatic machine fault diagnosis method using Spectrogram images. The proposed method uses STFT to effectively extract feature vectors from frequencies generated by machine defects, and the feature vectors detected by STFT were converted into spectrogram images and classified by CNN by machine status. The results show that the proposed method can be effectively used not only to detect defects but also to various automatic diagnosis system based on sound.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.21 no.12
• /
• pp.107-114
• /
• 2020

Technologies related scan-to-BIM for BIM-based reverse engineering techniques are beginning to be actively introduced in the A.E.C. industry, and the scalability of the technology is growing considerably. This study uses patent analysis based on objective data to find the right direction for Korean Scan-to-BIM technology by identifying the trends in Korea, the United States, Europe, and Japan. This was done using the WIPSON patent search system to find previous research on patent analysis related to building technology, theoretical consideration of scan-to-BIM technology, and published patents. We collected information, verified the process, and extracted valid patents. We used the effective patent data to analyze the annual trend of patent applications, national trends, and technological trends through the International Patent Classification (IPC) code, the types of the top 20 major applicants, and family patent trends.

• Industry Promotion Research
• /
• v.7 no.1
• /
• pp.1-8
• /
• 2022

We investigated various types of machine learning methods that can be applied to censored data. Exploratory data analysis reveals the distribution of each feature, relationships among features. Next, classification problem has been set up where the dependent variable is death_event while the rest of the features are independent variables. After applying various machine learning methods to the data, it has been found that just like many other reports from the artificial intelligence arena random forest performs better than logistic regression. But recently well performed artificial neural network and gradient boost do not perform as expected due to the lack of data. Finally Kaplan-Meier and Cox proportional hazard model have been employed to explore the relationship of the dependent variable (t_i, δ_i) with the independent variables. Also random forest which is used in machine learning has been applied to the survival analysis with censored data.

• Animal Bioscience
• /
• v.35 no.6
• /
• pp.791-803
• /
• 2022

Genome/gene-editing (GE) techniques, characterized by a low technological barrier, high efficiency, and broad application among organisms, are now being employed not only in medical science but also in agriculture/veterinary science. Different engineered CRISPR/Cas9s have been identified to expand the application of this technology. In pig production, GE is a precise new breeding technology (NBT), and promising outcomes in improving economic traits, such as growth, lean or healthy meat production, animal welfare, and disease resistance, have already been documented and reviewed. These promising achievements in porcine gene editing, including the Myostatin gene knockout (KO) in indigenous breeds to improve lean meat production, the uncoupling protein 1 (UCP1) gene knock-in to enhance piglet thermogenesis and survival under cold stress, the generation of GGTA1 and CMP-N-glycolylneuraminic acid hydroxylase (CMAH) gene double KO (dKO) pigs to produce healthy red meat, and the KO or deletion of exon 7 of the CD163 gene to confer resistance to porcine reproductive and respiratory syndrome virus infection, are described in the present article. Other related approaches for such purposes are also discussed. The current trend of global regulations or legislation for GE organisms is that they are exempted from classification as genetically modified organisms (GMOs) if no exogenes are integrated into the genome, according to product-based and not process-based methods. Moreover, an updated case study in the EU showed that current GMO legislation is not fit for purpose in term of NBTs, which contribute to the objectives of the EU's Green Deal and biodiversity strategies and even meet the United Nations' sustainable development goals for a more resilient and sustainable agri-food system. The GE pigs generated via NBT will be exempted from classification as GMOs, and their global valorization and commercialization can be foreseen.

• Journal of the Semiconductor & Display Technology
• /
• v.22 no.1
• /
• pp.72-77
• /
• 2023

In many manufacturing settings, including the semiconductor industry, products are completed by producing and assembling various components. Sorting out from randomly mixed parts and classification operations takes a lot of time and labor. Recently, many efforts have been made to select and assemble correct parts from mixed parts using robots. Automating the sorting and classification of randomly mixed components is difficult since various objects and the positions and attitudes of robots and cameras in 3D space need to be known. Previously, only objects in specific positions were grasped by robots or people sorting items directly. To enable robots to pick up random objects in 3D space, bin picking technology is required. To realize bin picking technology, it is essential to understand the coordinate system information between the robot, the grasping target object, and the camera. Calibration work to understand the coordinate system information between them is necessary to grasp the object recognized by the camera. It is difficult to restore the depth value of 2D images when 3D restoration is performed, which is necessary for bin picking technology. In this paper, we propose to use depth information of RGB-D camera for Z value in rotation and movement conversion used in calibration. Proceed with camera calibration for accurate coordinate system conversion of objects in 2D images, and proceed with calibration of robot and camera. We proved the effectiveness of the proposed method through accuracy evaluations for camera calibration and calibration between robots and cameras.

• Industry Promotion Research
• /
• v.8 no.2
• /
• pp.141-148
• /
• 2023

The purpose of this study is to analyze NSSI (Non-suicidal Self-Injury) and upcoming papers to review research achievements and endpoints and current problems. Analysis data necessary for the study were collected through the classification procedure as domestic suspension during the mandatory period related to self-harm from 2010 to the present. As a result of analysis according to classification categories (general characteristics, study subjects, research variables, program utilization), first, domestic studies related to non-suicidal self-harm increased steadily through 2019 and 2022 after increasing in 2017 (three articles) showing shape. Second, changes in the characteristics of the research subjects began to appear. If previous studies had mainly focused on adolescents who experienced self-harm, it was expanded to early adulthood (college students), and the subjects of the study diversified to include those who had experienced self-harm cessation, counselors, and parents. Third, various research methods began to appear. Compared to 2017, when quantitative research was active, research was conducted that applied various qualitative research methods (narrative, phenomenology, grounded theory, meta-analysis, case study). Finally, discussion of the research results and suggestions for future research were added.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.11
• /
• pp.1571-1576
• /
• 2022

This study aims to analyze the interest of modern people in non-face-to-face medical counseling in the medical industrys. Big data was collected on two social platforms, 지식인, a platform that allows experts to receive medical counseling, and YouTube. In addition to the top five keywords of telephone counseling, "internal medicine", "general medicine", "department of neurology", "department of mental health", and "pediatrics", a data set was built from each platform with a total of eight search terms: "specialist", "medical counseling", and "health information". Afterwards, pre-processing processes such as morpheme classification, disease extraction, and normalization were performed based on the crawled data. Data was visualized with word clouds, broken line graphs, quarterly graphs, and bar graphs by disease frequency based on word frequency. An emotional classification model was constructed only for YouTube data, and the performance of GRU and BERT-based models was compared.