Earticle

코로나19의 확산과 함께 야외 운동이 제한되면서 실내에서 체감형 게임을 이용하는 이용자 수가 늘고 있다. 게이머를 대신하는 아바타 형식의 캐릭터 디자인 커스터마이징 시스템은 게이머에게 게임에 대한 매력을 높일 수 있다는 측면에서 개발자들에게 중요한 연구주제의 하나이다. 본 논문은 판매량이 많고 캐릭터 커스터마이징이 가능 한 체감형 게임을 게임 분석 사례로 선택하였다. 선택된 세 개 게임의 캐릭터 커스터마이징 시스템 분석을 통해 커스터마이징 시스템의 중요한 요소를 추출하였다. 마지막으로 체감형 게임 캐릭터 커스터마이징 시스템의 단점을 도출하고 선택 시스템，변형 시스템，부분 시스템에 대해 개선안을 제시하였다. 이러한 제안이 향후 체감형 게임 캐릭터 커스터마이징 시스템 설계에 나름의 참고가 되기를 기대한다.

With the spread of COVID-19 and restrictions on outdoor exercise, the number of users using motion sensing games indoors is increasing. An avatar-type character design customization system on behalf of gamers is one of the important research topics for developers in that it can increase the game's appeal to gamers. In this paper, a motion sensing game with high sales volume and character customization is selected as a game analysis case. Through the analysis of the character customization system of the three selected games, important elements of the customization system were extracted. Finally, the shortcomings of the motion sensing game character customization system were derived, and improvement plans were presented for the selection system, transformation system, and partial system. It is expected that this proposal will serve as a reference for designing a motion sensing game character customization system in the future.

Rapid progress and widespread outbreak of COVID-19 have caused devastating influence on the health systems all around the world. The importance of countermeasures to tackle this problem lead to widespread use of Computer Aided Diagnosis (CADs) applications using deep neural networks. The unprecedented success of machine learning techniques, especially deep learning networks in medical images, have led to their recent prominence in improving efficient diagnosis of COVID-19 with increased detection accuracy. However, recent studies in the field of security of AI-based systems revealed that these deep learning models are vulnerable to adversarial attacks. Adversarial examples generated by attack algorithms are not recognizable by the human eye and can easily deceive the state-of-the-art deep learning models, therefore they threaten security-critical learning applications. In this paper, the methodology, results and concerns of recent works on robustness of AI based COVID-19 systems are summarized and discussed. We explore important security concerns related to deep neural networks and review current state-of-the-art defense methods to prevent performance degradation.

Background An artificial intelligence-assisted motion-analysis system without markers, OpenPose, is used to calculate joint angles in sports and medical analyses. Measuring the angles of the hip and knee joint in the frontal plane during the performance of weight-bearing activities is valuable in patients with patellofemoral pain syndrome (PFPS). Purpose The purpose of this study was to assess the validity of OpenPose using a pre-trained human motion-tracking algorithm for measuring the angles of the hip and knee joint in the frontal plane during standing hip abduction, semi-squat movements, and forward step-down movements compared with marker-based three-dimensional motion analysis. Study design Cross-sectional study Methods Eight individuals with PFPS participated in the current study. To investigate the validity of OpenPose, the angles of the hip and knee in the frontal plane were measured simultaneously with a smartphone camera using the OpenPose library and Vicon as the gold-standard motion-analysis system while performing three weight-bearing activities. Pearson and Spearman correlation analysis was used to assess the validity of the OpenPose-based motion-analysis system. Results Correlation coefficients ranged from 0.04 to 0.61 on the more symptomatic side and from 0.02 to 0.88 on the less symptomatic side for the three weight-bearing activities. When performing standing hip abduction and step-down movements, the validity of the measurements of hip abduction was fair or good. When performing semi-squat movements, the validity of the knee abduction measurements was fair. Conclusions The OpenPose-based motion-analysis system can provide fair or good level of validity of measurements of frontal hip and knee angles during weight-bearing activities of individuals with PFPS in real environments and for remote rehabilitation.

Objective To investigate the differences in biomechanical parameters measured by gait analysis systems between healthy subjects and subjects with plantar fasciitis (PF), and to compare biomechanical parameters between ‘normal, barefooted’ gait and arch building gait in the participants.Methods The researchers evaluated 15 subjects (30 feet) with bilateral foot pain and 15 subjects (15 feet) with unilateral foot pain who had a clinical diagnosis of PF. Additionally, 17 subjects (34 feet) who had no heel pain were recruited. Subjects were excluded if they had a traumatic event, prior surgery or fractures of the lower limbs, a leg length discrepancy of 1 cm or greater, a body mass index greater than 35 kg/m2, or had musculoskeletal disorders. The participants were asked to walk with an arch building gait on a treadmill at 2.3 km/hr for 5 minutes. Various gait parameters were measured.Results With the arch building gait, the PF group proved that gait line length and single support line were significantly decreased, and lateral symmetry of the PF group was increased compared to that of the control group. The subjects with bilateral PF displayed significantly increased maximum pressure over the heel and the forefoot during arch building gait. In addition, the subjects with unilateral PF showed significantly increased maximum pressure over the forefoot with arch building gait.Conclusion The researchers show that various biomechanical differences exist between healthy subjects and those with PF. Employing an arch building gait in patients with PF could be helpful in changing gait patterns to normal biomechanics.

Objective To evaluate the gait pattern of patients with gait disturbances without consideration of defilades due to assistive devices. This study focuses on gait analysis using the inertial measurement unit (IMU) system, which can also be used to determine the most appropriate assistive device for patients with gait disturbances. Methods Records of 18 disabled patients who visited the Department of Rehabilitation from May 2018 to June 2018 were selected. Patients’ gait patterns were analyzed using the IMU system with different assistive devices to determine the most appropriate device depending on the patient’s condition. Evaluation was performed using two or more devices, and the appropriate device was selected by comparing the 14 parameters of gait evaluation. The device showing measurements nearer or the nearest to the normative value was selected for rehabilitation. Results The result of the gait evaluation in all 18 patients was analyzed using the IMU system. According to the records, the patients were evaluated using various assistive devices without consideration of defilades. Moreover, this gait analysis was effective in determining the most appropriate device for each patient. Increased gait cycle time and swing phase and decreased stance phase were observed in devices requiring significant assistance. Conclusion The IMU-based gait analysis system is beneficial in evaluating gait in clinical fields. Specifically, it is useful in evaluating patients with gait disturbances who require assistive devices. Furthermore, it allows the establishment of an evidence-based decision for the most appropriate assistive walking devices for patients with gait disturbances.

Background Standardized clinical terminology is essential for semantic interoperability and for effectively utilizing clinical data in digital health, research, and artificial intelligence (AI). Although manual therapy is widely used in physical therapy practice, its techniques have not yet been systematically integrated into major standardized terminology systems such as systematized nomenclature of medicine clinical terms (SNOMED CT) and observational medical outcomes partnership common data model (OMOP CDM). Purpose The purpose of this study was to evaluate the status of terminology standardization for manual therapy techniques within SNOMED CT and OMOP CDM, including the identification of alternative or broader concepts when direct representations were unavailable. Study design Terminology evaluation study Methods Twenty-nine techniques were selected through a literature review, educational resources, and expert consensus. Each technique was evaluated for direct representation in SNOMED CT and for mapping within the OMOP CDM. Techniques not directly defined were assessed for the availability of broader, more generic concepts. Results Among the 29 techniques, 15 (51.7%) were standardized in at least one system, while 14 (48.3%) were not standardized and could only be described using higher-level categories. The standardization rates were 40.0% for orthopedic techniques, 44.4% for soft tissue techniques, 60.0% for nerve techniques, and 80.0% for other integrative approaches. Conclusions Nearly half of the manual therapy techniques commonly used in physical therapy lack independent representation in SNOMED CT and OMOP CDM. Although generic categories allow for minimal documentation, the absence of detailed, granular concepts diminishes semantic accuracy, impedes cross-institutional comparability, and limits multicenter research and AI-based applications.

The aim of this study was to investigate the effects of heat stress on milk traits in South Korea using comprehensive data (dairy production and climate). The dataset for this study comprised 1,498,232 test-day records for milk yield, fat- and protein-corrected milk, fat yield, protein yield, milk urea nitrogen (MUN), and somatic cell score (SCS) from 215,276 Holstein cows (primiparous: n = 122,087; multiparous: n = 93,189) in 2,419 South Korean dairy herds. Data were collected from July 2017 to April 2020 through the Dairy Cattle Improvement Program, and merged with meteorological data from 600 automatic weather stations through the Korea Meteorological Administration. The segmented regression model was used to estimate the effects of the temperature-humidity index (THI) on milk traits and elucidate the break point (BP) of the THI. To acquire the least-squares mean of milk traits, the generalized linear model was applied using fixed effects (region, calving year, calving month, parity, days in milk, and THI). For all parameters, the BP of THI was observed; in particular, milk production parameters dramatically decreased after a specific BP of THI (p < 0.05). In contrast, MUN and SCS drastically increased when THI exceeded BP in all cows (p < 0.05) and primiparous cows (p < 0.05), respectively. Dairy cows in South Korea exhibited negative effects on milk traits (decrease in milk performance, increase in MUN, and SCS) when the THI exceeded 70; therefore, detailed feeding management is required to prevent heat stress in dairy cows.

본 연구는 2011년 도입된 진로진학상담교사 제도의 형성 과정과 운영실태를 체제 분석 모형을 통해 종합적으로 분석하고, 향후 정책 개선 방향을 제시하는 데 목적이 있다. 연구 방법으로는 정책 문서 분석, 교육통계 분석, 경기도 진로교사 대상 설문조사(n=327)를 실시하였다. 경기도를 연구 지역으로 선정한 이유는 전국 최대 규모의 진로교사 배치 지역으로서 제도 운영의 특성이 가장 잘 드러나며, 다양한 학교 유형과 지역특성을 포괄하기 때문이다. 혼합연구방법(mixed methods)을 채택한 이유는 정책 분석이라는 거시적 관점과 현장 교사의 실천 경험이라는 미시적 관점을 통합함으로써, 제도의 형성 맥락과 운영 실태를 입체적으로파악할 수 있기 때문이다. 분석 결과, 투입 측면에서는 법적·제도적 기반이 마련되었으나 예산 및 인력 배치의 지역 간 격차가 확인되었다. 과정측면에서는 진로상담 및 진로수업 운영은 활성화되었으나, 과중한 행정업무로 인한 본연의 역할 수행 제약이 나타났다. 산출 측면에서는 학생의진로결정 자기효능감 향상 등 긍정적 성과가 있었으나(다만 이는 한국직업능력연구원의 외부 조사를 재인용한 것으로, 진로교사 배치의 순수 효과를 인과적으로 규명한 것이 아님을 밝힌다), 학교급 간 연계 부족과전문성 개발 체계의 미흡함이 한계로 지적되었다. 이를 토대로 본 연구는전문성 강화를 위한 생애주기별 재교육 체계 구축, 행정 업무 경감을 위한 거버넌스 확립, 학생 수 기반 복수 배치 및 초등학교 배치 의무화 등의정책적 개선 과제를 제안하였다. 단, 정책 제언의 단계적 실행 전략과우선순위에 대한 논의도 함께 제시하였다.

The purpose of this study is to comprehensively analyze the formation process and operation of the career counseling teacher system introduced in 2011 through a system analysis model, and to suggest future policy improvement directions. As research methods, policy document analysis, educational statistics analysis, and questionnaire survey for career teachers in Gyeonggi-do (n=327) were conducted. The reason why Gyeonggi-do was selected as a research area is that the characteristics of the system operation are best revealed as the largest career teacher placement area in the country, and it encompasses various school types and regional characteristics. The reason for adopting mixed methods is that by integrating the macroscopic view of policy analysis and the microscopic view of the practical experience of field teachers, it is possible to grasp the context of formation and operation of the system in three dimensions. As a result of the analysis, the legal and institutional basis was established in terms of input, but the gap between regions in budget and manpower allocation was confirmed. In terms of process, career counseling and career class operation were activated, but restrictions on the performance of the original role were found due to heavy administrative work. In terms of output, there were positive achievements such as improving students' career decision-making self-efficacy (but this is a re-citation of an external survey by the Korea Vocational Competency Institute, and it is revealed that the pure effect of career teacher placement is not causally identified), but the lack of linkage between school levels and lack of professional development system were pointed out as limitations. Based on this, this study proposed policy improvement tasks such as establishing a re-education system for each life cycle to strengthen professionalism, establishing governance to reduce administrative work, multiple placement based on the number of students, and mandatory elementary school placement. However, discussions on step-by-step implementation strategies and priorities of policy suggestions were also presented.

As a new paradigm, semantic communication (SC) is attracting attention as the next-generation communication. In order to develop semantic communication systems, the research trend is to adopt deep learning networks with the best performance in each field of deep learning. Deep-learning based SC systems are task-dependent because the deep networks employed are specialized to perform specific tasks. Therefore, SC systems cannot be used for users using applications that do not work with the deep networks exploited by the SC systems. Such users still should be served by a traditional communication (TC) system. Hence, in the near future, communication networks will face the coexistence of both SC system and TC system, and thus a base station (BS) has to simultaneously transmit symbol streams encoded by an SC system and a TC system via orthogonal frequency-division multiple access. Although the current works verify that SC systems outperform a TC system in a point-to-point communication, it is necessary to analyze whether the introduction of SC systems can also improve the network performance provided by a BS. In this paper, we investigate the impact of introducing the new emerging semantic communication on network performance by modeling signal-to-noise ratio of SC users served by SC systems and analyzing sum-rate maximization with a minimum required SNR constraint for SC users. Via extensive numerical results, changes in network performance with and without an SC system are discussed with varying the distances from a BS, the number of SC users served by an SC system, and the degree of SC performance.

In this paper, we investigate the performance of a downlink non-orthogonal multiple access (NOMA) based visible light communication (VLC) system. We consider that the users are randomly located following a binary Poisson point (BPP) Process. The performance of the considered system is evaluated by deriving novel closed-form expressions of the bit error ratio (BER) and ergodic sum rate. The impact of an error due to the successive interference cancellation (SIC) is also taken into consideration in the BER analysis. The obtained results with respect to different system parameters are instrumental to provide useful insights into the system performance.

Numerical study on a lifting system developed for shallow sea test is performed by a CFD model. In this study, the flow behaviors in a lifting pump and flexible hoses is simulated by using FLUENT, a commercial CFD model. Two cases for calculation the pump cavitation behavior are selected according to the hose length and the depth of lifting pump. From the numerical results, it is found that the vortex flow is observed in the first-stage stator and the calculated NPSH meets the minimum required by a lifting pump based on the shallow sea environments.

Filter bank multiple carrier (FBMC) technology is one of the alternative solutions for multicarrier modulation. FBMC does not need cyclic prefix (CP) and guard band utilized for orthogonal frequency division multiplexing (OFDM) and CP and guard band cause loss of spectral efficiency. FBMC features offset QAM (OQAM) and band-limited filtering on each subcarrier, which eliminate the need of CP and guard band. FBMC filtering could maintain the orthogonality in a real signal domain by using well localized filter. Meanwhile layer division multiplexing (LDM) is also introduced to increase the spectral efficiency. In LDM, low density parity check (LDPC) coded multi signals are transmitted simultaneously with different power levels in same frequency band and these signals form signal layers. Combination of FBMC and LDM techniques can maximize the spectral efficiency. In this paper, LDM system which adopts FBMC is proposed. The LDM system has both OFDM and FBMC modulated layers. To apply FBMC to LDM system, log-likelihood ratio (LLR) calculation scheme for FBMC is needed for LDPC decoding. Three scenarios for LDM system are considered and BER performance of each scenario is analyzed to find proper scenario.

최근 중국의 3대 온라인 플랫폼, 바이두, 알리바바, 텐센트는 예전에 겪어보 지 못했던 다양한 어려움에 직면하고 있다. 특히 개인 정보의 부적절한 사용, 중국 청소년들의 인터넷 중독, 모조품 유통과 거래 및 비윤리적인 온라인 마케 팅 관행 등은 3대 온라인 플랫폼의 저조한 국제화 실적과 깊은 관련이 있다. 무엇보다도 2015년 이후, 상승세가 크게 둔화된 바이두의 매출액 추이에 좀 더 주목할 필요가 있는데, 시간이 지날수록 바이두와 알리바바 및 텐센트 사이의 매출액 차이가 점점 커지고 있다. 이에 본 연구는 바이두 경영성과의 부진 이 유를 전략적 포지셔닝 관점에서 분석하였다. 일반적으로 기업의 전략적 포지셔 닝은 다양한 제품과 서비스 중심 활동, 소비자 니즈(needs) 중심 활동, 소비자 접근성 중심 활동 등 세 가지 유형의 활동에서 형성된다. 따라서 본 연구는 바 이두, 알리바바, 텐센트의 기업활동들을 세 가지 유형으로 재구성하고, 전체적 인 활동 시스템 지도를 작성함으로써, 어떠한 유형의 활동들이 다른 유형의 활 동들보다 더 많은지, 즉 어떠한 종류의 포지셔닝으로 볼 수 있는지를 분석하였 다. 그 결과 바이두의 경우, 많은 활동들이 서비스 다양성 중심의 포지셔닝에 집중된 것으로 나타났다. 더 나아가 바이두의 소비자 니즈(needs) 중심 활동들 은 사용자들에게 독특한 가치를 제공하는데 있어서, 알리바바나 텐센트에 비해 상대적으로 빈약한 것으로 나타났다.

Recently the leadership of Chinese big three online firms, Baidu, Alibaba, Tencent, called the BAT confronts with various difficulties that rarely managed before. The difficulties include the personal data misuse, internet addition, counterfeit goods trade and unethical online marketing, and they work as the direct reason for poor performances of their globalization. In particular, the revenue of Baidu began to be clearly sluggish from 2015, so the revenue gap between Baidu and other two rivals largely increased from that time. This study analyzed the reason for Baidu’s recent downturn from the strategic positioning perspective. The strategic positioning of a firm can be made by three types of business activities; customer needs-based activity, access-based activity, product or service variety-based activity. This study reorganized the business activities of BAT, and displayed the activity system of each firm in order to decide which type of business activities is relatively more than other types. In result, many activities of Baidu turned out to be more oriented for variety-based positioning rather than needs-based positioning or access-based one. Further the needs-based type of activities of Baidu should have been more specific to create a distinctive value for users.

최근 기후변화로 홍수, 국지성호우, 태풍 등의 재해가 매년 발생하고 있다. 과거 피해가 빈번히 발생하는 치수 위주의 연구가 진행되었으나, 최근 도시물순환에 대한 관심이 높아지면서 물순환 관리에 대한 인식이 증가하고 있다. 물순환을 관리하는 가장 우선되어야 하는 것은 정량적인 물관리 목표를 설정과 정확한 물순환 적용지점을 설정하는 것이라 할 수 있다. 이에 본 연구에서는 기존 도시유역에서 토지피복도만을 활용한 물순환 분석의 한계점을 개선하고자 중첩지도 분석을 실시하여 토지피복도로 분석하지 못한 지점 을 건물면적과 건물정보가 통합된 분석을 실시하였다. LID 시설물의 효과 검토를 위해 모델을 활용하여 서울특별시 효자동 지역을 대상으로 분석하였으며, 그 결과 재현기간 5년, 10년 강우지속기간 10분에서 첨두홍수량은 28%, 30% 저감, 유출체적은 18%, 14% 저감, 침수면적은 92%, 93%가 저감된 것으로 분석되 었다.

Recently, disasters such as floods, localized heavy rains, and typhoons have occurred every year due to climate change. In the past, studies focused on flood control, where damage frequently occurred, but recently, as interest in urban water circulation has increased, awareness of water circulation management has increased. The most important thing to manage the water cycle is to set quantitative water management goals and the exact water cycle application point. Therefore, in this study, to improve the limitations of the water cycle analysis using only the land cover map in the existing urban watershed, an overlapping map analysis was conducted, and an integrated analysis of the building area and building information was conducted for the points where the land cover map was not analyzed. To review the effectiveness of LID facilities, the model was used to analyze Hyoja-dong, Seoul. It was analyzed that peak flood volume was reduced by 28% and 30%, the Return period 5 years, 10 years Rain duration 10min runoff volume was reduced by 18% and 14%, and the flooded area was reduced by 92% and 93%.

농업용수는 우리나라 수자원 중 60%이상을 차지하는 아주 중요한 수자원이다. 농업용수의 효율적인 관리를 위해서 유역전반에 걸친 물수급 현황, 유역 내 각 시설물의 용수공급능력 분석, 유역 내 물부 족 여부 및 추가용수개발 가능량 분석 등 용수이용의 효율성 제고를 위한 자료를 분석할 수 있는 시스템이 동반되어야 한다. 본 연구로 개발된 WBSIr은 유출모형의 매개변수 추정에 있어 다목적 유전자 알고리즘을 적용하여 최적화하고 실측 유출량의 확보가 가능한 경우 각 소유역마다 TANK 모형의 적용이 가능하다. WBSIr을 적용하여 유역의 관개용수 공급 네트워크를 고려하여 물수지를 분석한 결과, 기존의 방법에 비해 보다 효율적인 소유역별 용수배분이 가능하였으며 취입보, 양수장 등 용수공급 시설물의 연계분석으로 유역전체의 통합적인 관리가 가능하게 되었다. 기존의 물수지 방법을 적용한 경우와 비교하였으며, 용수공급량 등의 산정이 보다 신뢰도 높은 결과를 나타냄으로 써 관개용수 공급 네트워크를 고려한 물수지 분석 수행이 타당함을 제시하였다.

Agricultural water accounts for more than 60 percent of water resources in South Korea. The efficient management of agricultural water requires a system that analyzes the data to improve water efficiency, including water supply and demand throughout the watershed, water supply capacity of each facility in the watershed, water shortage in the watershed, and the possibility of developing additional water supplies. WBSIr, developed in this study, is optimized by applying a multi-purpose genetic algorithm in estimating parameters of the spill model, and the TANK model can be applied to each watershed if actual runoff is measured. Based on the WBSIr analysis of water balance considering an irrigation water supply network in the basin, it is possible to allocate water to each watershed more efficiently than the previous method. In comparison to the conventional water balance method, the water balance analysis considering the irrigation water supply network was found more reliable in the assessment of water supply.

It is crucial to develop effective and efficient big data analytics methods for problem solving in the field of business in order to improve the performance of data analytics and reduce costs and risks in the analysis of customer data. In this study, a big data-driven data analysis system using artificial intelligence techniques is designed to increase the accuracy of big data analytics along with the rapid growth of the field of data science. We present a key direction for big data analysis systems through missing value imputation, outlier detection, feature extraction, utilization of explainable artificial intelligence techniques, and exploratory data analysis. Our objective is not only to develop big data analysis techniques with complex structures of business data but also to bridge the gap between the theoretical ideas in artificial intelligence methods and the analysis of real-world data in the field of business.

과거의 자연재해 대응은 주로 상황근무자의 재난 대응경험과 주관적인 판단에 의한 대응이 이루어 져왔다. 하지만 최근에는 전에 경험할 수 없던 재난들이 발생하면서 시스템을 통한 예측 및 대응 등의 재난관리가 요구되고 있으며, 시스템의 물리적 모델링을 통한 피해 예측의 불확실성을 줄이기 위해 과거 피해 정보를 포함한 재해 통계를 이용한 재난대응의 중요성도 대두되고 있다. 따라서 본 연구에서는 우리나라에서 자연재난 피해를 빈번하게 유발시키는 호우와 태풍을 중심으로 국가 재해통계 기반의 재해상황 분석 시스템을 개발하였다. 시스템의 주요 기능은 호우․태풍 특보지역에 대한 과거 재해피해 이력, 자료현황 통계 제공 및 유사태풍 조회로 구성하였다. 시스템은 현재 호우․ 태풍 발생 시 중앙재난안전대책본부의 재난대응 업무에서 활용되고 있으며, 향후 국가 차원의 자연 재난 대응업무를 지원하기 위해서는 호우 및 태풍 외에도 강풍, 대설 등의 모든 자연재난에 활용할 수 있도록 시스템을 확장해야 할 것으로 사료된다.

The conventional disaster response in South Korea was mainly based on experiences and subjective judgement of on-site responders. However, a growing number of new types of disasters requires a systematic disaster response and management. In order to reduce the uncertainty of a physical modeling in damage prediction, the use of disaster statistics including the past damage data becomes more critical. Therefore, in this study, we developed a Disaster Situation Analysis System (DSAS) based on national disaster statistics on heavy rains and typhoons which frequently caused damages across South Korea. The main function of the system is to search the past history of disaster damage in the disaster areas and relevant data and statistics relating to similar cases. This system is currently being used for disaster response tasks of the Central Headquarters of Safety and Countermeasures Against Disasters in South Korea. In order to better cope with future disasters at the national level, the system used for heavy rains and typhoons should be expanded to include all types of natural disasters such as heavy snow and strong winds.

This research proposes the development of an online psychological counseling platform utilizing an Artificial Intelligence (AI)-based emotion analysis system. The platform, leveraging facial video, voice, and text data, aims to real-time identify and analyze the emotional states of counselees in a non-face-to-face counseling environment, providing counselors with the necessary information to facilitate appropriate counseling. The outcomes of this research are expected to enhance communication effectiveness between counselors and counselees and contribute to the psychological well-being of counselees in online counseling scenarios.