• Journal of the Korea Computer Graphics Society
• /
• v.30 no.1
• /
• pp.1-10
• /
• 2024

In immersive virtual environments including mixed reality (MR) and virtual reality (VR), avatars or agents, which are virtual humans, are being studied and applied in various ways as factors that increase users' social presence. Recently, studies are being conducted to apply generative AI as an agent to improve user learning effects or suggest a collaborative environment in an immersive virtual environment. This study proposes a novel method for interface application of a virtual assistant agent (VAA) using OpenAI's ChatGPT in an immersive virtual environment including VR and MR. The proposed method consists of an information agent that responds to user queries and a control agent that controls virtual objects and environments according to user needs. We set up a development environment that integrates the Unity 3D engine, OpenAI, and packages and development tools for user participation in MR and VR. Additionally, we set up a workflow that leads from voice input to the creation of a question query to an answer query, or a control request query to a control script. Based on this, MR and VR experience environments were produced, and experiments to confirm the performance of VAA were divided into response time of information agent and accuracy of control agent. It was confirmed that the interface application of the proposed VAA can increase efficiency in simple and repetitive tasks along with user-friendly features. We present a novel direction for the interface application of an immersive virtual environment through the proposed VAA and clarify the discovered problems and limitations so far.

• The Journal of the Acoustical Society of Korea
• /
• v.26 no.6
• /
• pp.250-258
• /
• 2007

Current speech recognition technology s achieved high performance with the development of hardware devices, however it is insufficient for some applications where high reliability is required, such as voice control of powered wheelchairs for disabled persons. For the system which aims to operate powered wheelchairs safely by voice in real environment, we need to consider that non-voice commands such as user s coughing, breathing, and spark-like mechanical noise should be rejected and the wheelchair system need to recognize the speech commands affected by disability, which contains specific pronunciation speed and frequency. In this paper, we propose non-voice rejection method to perform voice/non-voice classification using both YIN based fundamental frequency(F0) extraction and reliability in preprocessing. We adopted a multi-template dictionary and acoustic modeling based speaker adaptation to cope with the pronunciation variation of inarticulately uttered speech. From the recognition tests conducted with the data collected in real environment, proposed YIN based fundamental extraction showed recall-precision rate of 95.1% better than that of 62% by cepstrum based method. Recognition test by a new system applied with multi-template dictionary and MAP adaptation also showed much higher accuracy of 99.5% than that of 78.6% by baseline system.

• Journal of Yeungnam Medical Science
• /
• v.40 no.4
• /
• pp.388-393
• /
• 2023

Background: Differentiating between bacterial and nonbacterial colitis remains a challenge. We aimed to evaluate the value of serum procalcitonin (PCT) and C-reactive protein (CRP) in differentiating between bacterial and nonbacterial colitis. Methods: Adult patients with three or more episodes of watery diarrhea and colitis symptoms within 14 days of a hospital visit were eligible for this study. The patients' stool pathogen polymerase chain reaction (PCR) testing results, serum PCT levels, and serum CRP levels were analyzed retrospectively. Patients were divided into bacterial and nonbacterial colitis groups according to their PCR. The laboratory data were compared between the two groups. The area under the receiver operating characteristic curve (AUC) was used to evaluate diagnostic accuracy. Results: In total, 636 patients were included; 186 in the bacterial colitis group and 450 in the nonbacterial colitis group. In the bacterial colitis group, Clostridium perfringens was the commonest pathogen (n=70), followed by Clostridium difficile toxin B (n=60). The AUC for PCT and CRP was 0.557 and 0.567, respectively, indicating poor discrimination. The sensitivity and specificity for diagnosing bacterial colitis were 54.8% and 52.6% for PCT, and 52.2% and 54.2% for CRP, respectively. Combining PCT and CRP measurements did not increase the discrimination performance (AUC, 0.522; 95% confidence interval, 0.474-0.571). Conclusion: Neither PCT nor CRP helped discriminate bacterial colitis from nonbacterial colitis.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.26 no.1
• /
• pp.53-62
• /
• 2024

In this study, we built a model to estimate frost occurrence in South Korea using single-input deep learning and multi-input deep learning. Meteorological factors used as learning data included minimum temperature, wind speed, relative humidity, cloud cover, and precipitation. As a result of statistical analysis for each factor on days when frost occurred and days when frost did not occur, significant differences were found. When evaluating the frost occurrence models based on single-input deep learning and multi-input deep learning model, the model using both GRU and MLP was highest accuracy at 0.8774 on average. As a result, it was found that frost occurrence model adopting multi-input deep learning improved performance more than using MLP, LSTM, GRU respectively.

• Journal of IKEEE
• /
• v.28 no.2
• /
• pp.150-157
• /
• 2024

In this paper, we propose a novel algorithm for predicting the number of apples on an apple tree using a deep learning-based object detection model and a polynomial regression model. Measuring the number of apples on an apple tree can be used to predict apple yield and to assess losses for determining agricultural disaster insurance payouts. To measure apple fruit load, we photographed the front and back sides of apple trees. We manually labeled the apples in the captured images to construct a dataset, which was then used to train a one-stage object detection CNN model. However, when apples on an apple tree are obscured by leaves, branches, or other parts of the tree, they may not be captured in images. Consequently, it becomes difficult for image recognition-based deep learning models to detect or infer the presence of these apples. To address this issue, we propose a two-stage inference process. In the first stage, we utilize an image-based deep learning model to count the number of apples in photos taken from both sides of the apple tree. In the second stage, we conduct a polynomial regression analysis, using the total apple count from the deep learning model as the independent variable, and the actual number of apples manually counted during an on-site visit to the orchard as the dependent variable. The performance evaluation of the two-stage inference system proposed in this paper showed an average accuracy of 90.98% in counting the number of apples on each apple tree. Therefore, the proposed method can significantly reduce the time and cost associated with manually counting apples. Furthermore, this approach has the potential to be widely adopted as a new foundational technology for fruit load estimation in related fields using deep learning.

• Journal of Cadastre & Land InformatiX
• /
• v.54 no.1
• /
• pp.19-32
• /
• 2024

This study examined the feasibility of image-based surveys by detecting objects in facilities and agricultural land using the YOLO algorithm based on drone images and comparing them with the land category by law. As a result of detecting objects through the YOLO algorithm, buildings showed a performance of detecting objects corresponding to 96.3% of the buildings provided in the existing digital map. In addition, the YOLO algorithm developed in this study detected 136 additional buildings that were not located in the digital map. Plastic greenhouses detected a total of 297 objects, but the detection rate was low for some plastic greenhouses for fruit trees. Also, agricultural land had the lowest detection rate. This result is because agricultural land has a larger area and irregular shape than buildings, so the accuracy is lower than buildings due to the inconsistency of training data. Therefore, segmentation detection, rather than box-shaped detection, is likely to be more effective for agricultural fields. Comparing the detected objects with the land category by law, it was analyzed that some buildings exist in agricultural and forest areas where it is difficult to locate buildings. It seems that it is necessary to link with administrative information to understand that these buildings are used illegally. Therefore, at the current level, it is possible to objectively determine the existence of buildings in fields where it is difficult to locate buildings.

• Knowledge Management Research
• /
• v.25 no.2
• /
• pp.19-47
• /
• 2024

Organizational ability to analyze and utilize data plays an important role in knowledge management and decision-making. This study aims to investigate the potential application of large language models in corporate data analysis. Focusing on the field of human resources, the research examines the data analysis capabilities of these models. Using the widely studied IBM HR dataset, the study reproduces machine learning-based employee turnover prediction analyses from previous research through ChatGPT and compares its predictive performance. Unlike past research methods that required advanced programming skills, ChatGPT-based machine learning data analysis, conducted through the analyst's natural language requests, offers the advantages of being much easier and faster. Moreover, its prediction accuracy was found to be competitive compared to previous studies. This suggests that large language models could serve as effective and practical alternatives in the field of corporate data analysis, which has traditionally demanded advanced programming capabilities. Furthermore, this approach is expected to contribute to the popularization of data analysis and the spread of data-driven decision-making (DDDM). The prompts used during the data analysis process and the program code generated by ChatGPT are also included in the appendix for verification, providing a foundation for future data analysis research using large language models.

• Smart Media Journal
• /
• v.12 no.11
• /
• pp.57-66
• /
• 2023

In this paper, we present a study on an AI-based system utilizing the CCTV system within city buses to predict the intentions of boarding and alighting passengers, with the aim of preventing accidents. The proposed system employs the YOLOv7 Pose model to detect passengers, while utilizing an LSTM model to predict intentions of tracked passengers. The system can be installed on the bus's CCTV terminals, allowing for real-time visual confirmation of passengers' intentions throughout driving. It also provides alerts to the driver, mitigating potential accidents during passenger transitions. Test results show accuracy rates of 0.81 for analyzing boarding intentions and 0.79 for predicting alighting intentions onboard. To ensure real-time performance, we verified that a minimum of 5 frames per second analysis is achievable in a GPU environment. his algorithm enhance the safety of passenger transitions during bus operations. In the future, with improved hardware specifications and abundant data collection, the system's expansion into various safety-related metrics is promising. This algorithm is anticipated to play a pivotal role in ensuring safety when autonomous driving becomes commercialized. Additionally, its applicability could extend to other modes of public transportation, such as subways and all forms of mass transit, contributing to the overall safety of public transportation systems.

• Information Systems Review
• /
• v.25 no.1
• /
• pp.93-110
• /
• 2023

Recently, attempts have been made to convert unstructured text into vectors and to analyze vast amounts of natural language for various purposes. In particular, the demand for analyzing texts in specialized domains is rapidly increasing. Therefore, studies are being conducted to analyze specialized and general-purpose documents simultaneously. To analyze specific terms with general terms, it is necessary to align the embedding space of the specific terms with the embedding space of the general terms. So far, attempts have been made to align the embedding of specific terms into the embedding space of general terms through a transformation matrix or mapping function. However, the linear transformation based on the transformation matrix showed a limitation in that it only works well in a local range. To overcome this limitation, various types of nonlinear vector alignment methods have been recently proposed. We propose a vector alignment model that matches the embedding space of specific terms to the embedding space of general terms through end-to-end learning that simultaneously learns the autoencoder and regression model. As a result of experiments with R&D documents in the "Healthcare" field, we confirmed the proposed methodology showed superior performance in terms of accuracy compared to the traditional model.

• Journal of the Korean Society of Radiology
• /
• v.82 no.1
• /
• pp.116-127
• /
• 2021

Purpose To evaluate the feasibility of texture analysis of gray-scale ultrasound (US) images for staging of hepatic fibrosis. Materials and Methods Altogether, 167 patients who had undergone routine US and laboratory tests for a fibrosis-4 (FIB-4) index were included. Texture parameters were measured using a dedicated in-house software. Regions of interest were placed in five different segments (3, 5, 6, 7, 8) for each patient. The FIB-4 index was used as the reference standard for hepatic fibrosis grade. Comparisons of the texture parameters between different fibrosis groups were performed with the Student's t-test or Mann-Whitney U-test. Diagnostic performance was evaluated by receiver operating curve analysis. Results The study population comprised of patients with no fibrosis (FIB-4 < 1.45, n = 50), mild fibrosis (1.45 ≤ FIB-4 ≤ 2.35, n = 37), moderate fibrosis (2.35 < FIB-4 ≤ 3.25, n = 27), and severe fibrosis (FIB-4 > 3.25, n = 53). Skewness in hepatic segment 5 showed a difference between patients with no fibrosis and mild fibrosis (0.2392 ± 0.3361, 0.4134 ± 0.3004, respectively, p = 0.0109). The area under the curve of skewness for discriminating patients with no fibrosis from those with mild fibrosis was 0.660 (95% confidence interval, 0.551-0.758), with an estimated accuracy, sensitivity, specificity of 64%, 87%, 48%, respectively. Conclusion A significant difference was observed regarding skewness in segment 5 between patients with no fibrosis and patients with mild fibrosis.