• Advances in Computational Design
• /
• 제4권2호
• /
• pp.73-90
• /
• 2019

The growth in the number of single-member households is a critical issue worldwide, especially among the elderly. For those living alone, who may be unaware of their health status or routines that could improve their health, a continuous healthcare monitoring system could provide valuable feedback. Assessing the performance adequacy of activities of daily living (ADL) can serve as a measure of an individual's health status; previous research has focused on determining a person's daily activities and extracting the most frequently performed behavioral patterns using camera recordings or wearable sensing techniques. However, existing methods used to extract common patterns of an occupant's activities in the home fail to address the spatio-temporal dimensions of human activities simultaneously. Though multiple sequence alignment (MSA) offers some advantages - such as inherent containment of the spatio-temporal data in sequence format, and rapid identification of hidden patterns - MSA has rarely been used to extract in-home ADL routines. This research proposes a method to extract a household occupant's ADL routines from a cumulative spatio-temporal data log of occupancy collected using a non-intrusive method (i.e., a tomographic motion detection system). The findings from an occupant's 28-day spatio-temporal activity log demonstrate the capacity of the proposed approach to identify routine patterns of an occupant's daily activities and to reveal the order, duration, and frequency of routine activities. Routine ADL patterns identified from the proposed approach are expected to provide a basis for detecting/evaluating abrupt or gradual changes of an occupant's ADL patterns that result from a physical or mental disorder, and can offer valuable information for home automation applications by enabling the prediction of ADL patterns.

• Advances in Computational Design
• /
• 제4권2호
• /
• pp.155-177
• /
• 2019

Operation and Maintenance (O&M) phase is the main contributor to the total lifecycle cost of a building. Previous studies have described that Building Information Models (BIM), if available with detailed asset information and their properties, can enable rapid troubleshooting and execution of O&M tasks by providing the required information of the facility. Despite the potential benefits, there is still rarely BIM with Mechanical, Electrical and Plumbing (MEP) assets and properties that are available for O&M. BIM is usually not in possession for existing buildings and generating BIM manually is a time-consuming process. Hence, there is a need for an automated approach that can reconstruct the MEP systems in BIM. Previous studies investigated automatic reconstruction of BIM using architectural drawings, structural drawings, or the combination with photos. But most of the previous studies are limited to reconstruct the architectural and structural components. Note that mechanical components in the building typically require more frequent maintenance than architectural or structural components. However, the building mechanical drawings are relatively more complex due to various type of symbols that are used to represent the mechanical systems. In order to address this challenge, this paper proposed a symbol recognition framework that can automatically recognize the different type of symbols in the building mechanical drawings. This study applied vector-based computer vision techniques to recognize the symbols and their properties (e.g., location, type, etc.) in two vector-based input documents: 2D drawings and the symbol description document. The framework not only enables recognizing and locating the mechanical component of interest for BIM reconstruction purpose but opens the possibility of merging the updated information into the current BIM in the future reducing the time of repeated manual creation of BIM after every renovation project.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제14권11호
• /
• pp.4290-4309
• /
• 2020

In recent years, advancements in machine learning capabilities have allowed it to see widespread adoption for tasks such as object detection, image classification, and anomaly detection. However, despite their promise, a limitation lies in the fact that a network's performance quality is based on the data which it receives. A well-trained network will still have poor performance if the subsequent data supplied to it contains artifacts, out of focus regions, or other visual distortions. Under normal circumstances, images of the same scene captured from differing points of focus, angles, or modalities must be separately analysed by the network, despite possibly containing overlapping information such as in the case of images of the same scene captured from different angles, or irrelevant information such as images captured from infrared sensors which can capture thermal information well but not topographical details. This factor can potentially add significantly to the computational time and resources required to utilize the network without providing any additional benefit. In this study, we plan to explore using image fusion techniques to assemble multiple images of the same scene into a single image that retains the most salient key features of the individual source images while discarding overlapping or irrelevant data that does not provide any benefit to the network. Utilizing this image fusion step before inputting a dataset into the network, the number of images would be significantly reduced with the potential to improve the classification performance accuracy by enhancing images while discarding irrelevant and overlapping regions.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제14권11호
• /
• pp.4268-4289
• /
• 2020

Advanced traffic signal timing method plays very important role in reducing road congestion and air pollution. Reinforcement learning is considered as superior approach to build traffic light timing scheme by many recent studies. It fulfills real adaptive control by the means of taking real-time traffic information as state, and adjusting traffic light scheme as action. However, existing works behave inefficient in complex intersections and they are lack of feasibility because most of them adopt traffic light scheme whose phase sequence is flexible. To address these issues, a novel adaptive traffic signal timing scheme is proposed. It's based on actor-critic reinforcement learning algorithm, and advanced techniques proximal policy optimization and generalized advantage estimation are integrated. In particular, a new kind of reward function and a simplified form of state representation are carefully defined, and they facilitate to improve the learning efficiency and reduce the computational complexity, respectively. Meanwhile, a fixed phase sequence signal scheme is derived, and constraint on the variations of successive phase durations is introduced, which enhances its feasibility and robustness in field applications. The proposed scheme is verified through field-data-based experiments in both medium and high traffic density scenarios. Simulation results exhibit remarkable improvement in traffic performance as well as the learning efficiency comparing with the existing reinforcement learning-based methods such as 3DQN and DDQN.

• 응용통계연구
• /
• 제35권2호
• /
• pp.177-188
• /
• 2022

일반적으로 support vector machine (SVM)은 높은 수준의 분류 정확도를 제공함으로써 다양한 분야의 분류분석에서 널리 사용되고 있다. 그러나 SVM은 최적화 계산식이 이차계획법(quadratic programming)으로 공식화되어 많은 계산 비용이 필요하므로 대용량 자료의 분류분석에는 그 사용이 제한된다. 또한 불균형 자료(imbalanced data)의 분류분석에서는 다수집단에 편향된 분류함수를 추정함으로써 대부분의 자료를 다수집단으로 분류하여 소수집단의 분류 정확도를 현저히 감소시키게 된다. 이러한 문제점들을 해결하기 위하여 본 논문에서는 다수집단을 분할(divide)하고, 소수집단을 과대추출(oversampling)하여 여러 분류함수들을 추정하고 이들을 통합(conquer)하는 DOC-SVM 분류기법을 제안한다. 제안한 DOC-SVM은 분할정복 알고리즘을 다수집단에 적용하여 SVM의 계산 효율을 향상시키고, 과대추출 알고리즘을 소수집단에 적용하여 SVM 분류함수의 편향을 줄이게 된다. 본 논문에서는 모의실험과 실제자료 분석을 통해 제안한 DOC-SVM의 효율적인 성능과 활용 가능성을 확인하였다.

• Nuclear Engineering and Technology
• /
• 제54권7호
• /
• pp.2599-2605
• /
• 2022

The Korea Retrospective Dosimetry network (KREDOS) performed an inter-laboratory comparison to confirm the harmonization and reliability of the results of retrospective dosimetry using mobile phone. The mobile phones were exposed to ¹⁹²Ir while attached to the human phantoms in the field experiment, and the exposure doses read by each laboratory were compared. This paper describes the reference dosimetry performed to present the reference values for inter-comparison and to obtain additional information about the dose distribution. Reference dosimetry included both measurement using LiF:Mg,Cu,Si and calculation via MCNP simulation to allow a comparison of doses obtained with the two different methodologies. When irradiating the phones, LiF elements were attached to the phones and phantoms and irradiated at the same time. The comparison results for the front of the phantoms were in good agreement, with an average relative difference of about 10%, while an average of about 16% relative difference occurred for the back and side of the phantom. The differences were attributed to the different characteristics of the physical and simulated phantoms, such as anatomical structure and constituent materials. Nevertheless, there was about 4% of under-estimation compared to measurements in the overall linear fitting, indicating the calculations were well matched to the measurements.

• 융합신호처리학회논문지
• /
• 제22권3호
• /
• pp.99-103
• /
• 2021

최근 CNN(Convolutional Neural Network)이 다양한 컴퓨터 비전 분야에서 우수한 성능으로 널리 사용되고 있다. 그러나 CNN은 계산 집약적이고 많은 메모리가 요구되어 한정적인 하드웨어 자원을 가지는 모바일이나 IoT(Internet of Things) 기기에 적용하기 어렵다. 이런 한계를 해결하기 위해, 기존의 학습된 모델의 성능을 최대한 유지하며 네트워크의 크기를 줄이는 인공신경망 경량화 연구가 진행되고 있다. 본 논문은 신경망 압축 기술 중 하나인 프루닝(Pruning)의 문턱값을 동적으로 조정하는 CNN 압축 기법을 제안한다. 프루닝될 가중치를 결정하는 문턱값을 실험적, 경험적으로 정하는 기존의 기술과 달리 정확도의 저하를 방지하는 최적의 문턱값을 동적으로 찾을 수 있으며, 경량화된 신경망을 얻는 시간을 단축할 수 있다. 제안 기법의 성능 검증을 위해 MNIST 데이터 셋을 사용하여 LeNet을 훈련시켰으며, 정확도 손실 없이 약 1.3 ~ 3배의 시간을 단축하여 경량화된 LeNet을 얻을 수 있었다.

• Journal of Platform Technology
• /
• 제11권3호
• /
• pp.45-55
• /
• 2023

본 논문은 한국 영화 산업의 의사 결정자들이 온라인상에서의 영화의 흥행을 극대화할 수 있도록 지원하는 데 도움을 주고자 역대 박스오피스 영화를 수집하여 영화를 유형별로 군집화하고, 유형별 온라인 박스오피스를 예측하는 모델을 제시한다. 이를 위해 먼저 다양한 특성을 고려하여 영화의 흥행 요인을 식별하고, 계산 효율성을 고려하여 특성 차원을 줄인다. 다음으로 영화의 유형을 체계적으로 분류하고, 유형별 온라인 박스오피스를 예측하며 흥행에 이바지한 요소를 분석한다. 이때, AutoML (Automated Machine Learning) 기법을 활용함으로써 다양한 기계학습 알고리즘을 자동으로 구성하고, 문제에 최적화된 알고리즘을 선택함으로써 여러 알고리즘을 쉽게 시도 및 선택한다. 이를 통해 정보화된 판단을 내릴 수 있는 기반을 제공하고, 영화 산업의 더 나은 성과를 도모하는 데 이바지할 것으로 기대할 수 있다.

• 산업과 과학
• /
• 제2권3호
• /
• pp.8-14
• /
• 2023

대다수의 IoT 기기들은 이미 AIoT를 사용하고 있지만, AI 애플리케이션을 구축하기 위해서는 아직 해결해야 할 문제가 많이 남아 있다. 본 연구에서는 IoT 에지 자원을 보다 효과적으로 분산하기 위해 머신러닝 기반의 IoT 에지 자원 관리 기법을 제안한다, 제안 기법은 머신러닝을 이용하여 IoT 에지 자원 동향을 파악함으로써 IoT 자원의 할당을 지속적으로 개선하며, 최적화된 IoT 자원은 머신러닝 컨볼루션을 활용하여 항상 변화하는 IoT 에지 자원을 안정적으로 유지한다, 제안 기법은 각각의 머신러닝 기반 IoT 에지 자원을 이전 패턴의 자원과 함께 해시값으로 저장함으로써 분산된 AIoT 맥락에서 공격 패턴으로 자원을 효과적으로 검증한다. 실험 결과에서는 IoT Edge 리소스의 무결성을 검증하기 위해서 이질적인 계산 하드웨어가 있는 복잡한 환경에서 잘 동작하는지 세 가지 다른 테스트 시나리오에서 에너지 효율성을 평가하였다.

• International Journal of Computer Science & Network Security
• /
• 제23권8호
• /
• pp.17-25
• /
• 2023

The alarming global prevalence of Type 2 Diabetes Mellitus (T2DM) has catalyzed an urgent need for robust, early diagnostic methodologies. This study unveils a pioneering approach to predicting T2DM, employing the Extreme Gradient Boosting (XGBoost) algorithm, renowned for its predictive accuracy and computational efficiency. The investigation harnesses a meticulously curated dataset of 4303 samples, extracted from a comprehensive Chinese research study, scrupulously aligned with the World Health Organization's indicators and standards. The dataset encapsulates a multifaceted spectrum of clinical, demographic, and lifestyle attributes. Through an intricate process of hyperparameter optimization, the XGBoost model exhibited an unparalleled best score, elucidating a distinctive combination of parameters such as a learning rate of 0.1, max depth of 3, 150 estimators, and specific colsample strategies. The model's validation accuracy of 0.957, coupled with a sensitivity of 0.9898 and specificity of 0.8897, underlines its robustness in classifying T2DM. A detailed analysis of the confusion matrix further substantiated the model's diagnostic prowess, with an F1-score of 0.9308, illustrating its balanced performance in true positive and negative classifications. The precision and recall metrics provided nuanced insights into the model's ability to minimize false predictions, thereby enhancing its clinical applicability. The research findings not only underline the remarkable efficacy of XGBoost in T2DM prediction but also contribute to the burgeoning field of machine learning applications in personalized healthcare. By elucidating a novel paradigm that accentuates the synergistic integration of multifaceted clinical parameters, this study fosters a promising avenue for precise early detection, risk stratification, and patient-centric intervention in diabetes care. The research serves as a beacon, inspiring further exploration and innovation in leveraging advanced analytical techniques for transformative impacts on predictive diagnostics and chronic disease management.