• 전자공학회논문지
• /
• 제50권6호
• /
• pp.254-259
• /
• 2013

중력 방향에 대한 가중치를 적용한 3축 가속도 센서 데이터를 낙상 특징 변수로 사용해서 은닉 마르코프 모델(Hidden Markov Model; HMM)에 적용한 새로운 낙상 인식 알고리듬을 제안한다. 기존에 낙상인식에 많이 사용되는 변수인 3축 가속도의 벡터 합(Sum Vector Magnitude, SVM)과 새롭게 정의한 변수인 중력방향가중치를 적용한 3축 가속도의 벡터 합(Gravity-weighted Sum Vector Magnitude, GSVM)를 포함한 다섯 가지 낙상특징변수를 은닉 마르코프 모델에 적용하여 낙상 인식률을 평가하였다. 실험을 통해 얻은 가장 좋은 결과는 중력방향가중치를 적용한 3축 가속도의 벡터 합 변수를 적용한 결과이고 100% 민감도(sensitivity)와 97.96% 특이성(specificity)를 얻었다. 이것은 단순 3축 가속도의 벡터 합 변수에 비해 민감도는 5.2%와 특이성은 4.5% 정도 향상되었다. 단순히 운동량만을 표현하는 3축 가속도의 벡터 합 변수에 비해 중력방향가중치를 적용한 3축 가속도의 벡터 합 변수가 낙상의 움직임에 대한 특징을 잘 표현하기 때문에 높은 인식률을 나타내었다.

• Journal of Multimedia Information System
• /
• 제7권4호
• /
• pp.257-262
• /
• 2020

In this paper, a method of detecting the emergency situations such as body fall is proposed by using color images. We detect body areas and key parts of a body through a pre-learned Mask R-CNN in the images captured by a camera. Then we find the centerline of the body through the joint points of both shoulders and feet. Also, we calculate an angle to the center line and then calculate the amount of change in the angle per hour. If the angle change is more than a certain value, then it is decided as a suspected fall. Also, if the suspected fall state persists for more than a certain frame, then it is determined as a fall situation. Simulation results show that the proposed method can detect body fall situation accurately.

• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2021년도 춘계학술대회
• /
• pp.552-554
• /
• 2021

도플러 레이더 센서로 취득한 움직임 데이터를 딥러닝 알고리듬을 사용한 낙상 인식 시스템을 제안한다. 딥러닝 알고리듬중 시계열 데이터에 장점을 가지는 RNN을 사용하여 낙상 인식에 적용한다. 도플러 레이더 센서의 낙상데이터는 시계열 데이터로 시간적인 특성을 가지고 있으며 결과는 낙상인지 아닌지 만을 판단하기 때문에 RNN의 구조를 시퀀스 입력에 고정 크기를 출력하는 구조로 설계하였다.

• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2021년도 추계학술대회
• /
• pp.226-228
• /
• 2021

구글사에서 출시된 ML Kit API의 Pose detection를 사용한 영상기반 낙상 알고리즘을 제안한다. Pose detection 알고리듬을 사용하여 추출된 신체의 33개의 3차원 특징점을 활용하여 낙상을 인식한다. 추출된 특징점을 분석하여 낙상을 인식하는 알고리듬은 k-NN을 사용한다. 영상의 크기와 영상내의 인체의 크기에 영향을 받지 않도록 정규화과정을 거치며 특징점들의 상대적인 움직임을 분석하여 낙상을 인식한다. 본 실험을 위해 사용한 13개의 테스트 영상중 13개의 영상에서 낙상을 인식하여 100%의 성공률을 보였다.

• 한국해양공학회지
• /
• 제35권1호
• /
• pp.91-97
• /
• 2021

Marine accidents caused by ships have brought about economic and social losses as well as human casualties. Most of these accidents are caused by small and medium-sized ships and are due to their poor conditions and insufficient equipment compared with larger vessels. Measures are quickly needed to improve the conditions. This paper discusses a video-integrated collision prediction and fall detection system to support the safe navigation of small- and medium-sized ships. The system predicts the collision of ships and detects falls by crew members using the CCTV, displays the analyzed integrated information using automatic identification system (AIS) messages, and provides alerts for the risks identified. The design consists of an object recognition algorithm, interface module, integrated display module, collision prediction and fall detection module, and an alarm management module. For the basic research, we implemented a deep learning algorithm to recognize the ship and crew from images, and an interface module to manage messages from AIS. To verify the implemented algorithm, we conducted tests using 120 images. Object recognition performance is calculated as mAP by comparing the pre-defined object with the object recognized through the algorithms. As results, the object recognition performance of the ship and the crew were approximately 50.44 mAP and 46.76 mAP each. The interface module showed that messages from the installed AIS were accurately converted according to the international standard. Therefore, we implemented an object recognition algorithm and interface module in the designed collision prediction and fall detection system and validated their usability with testing.

• 센서학회지
• /
• 제26권1호
• /
• pp.7-14
• /
• 2017

A threshold-based fall recognition algorithm using a tri-axial accelerometer and a bi-axial gyroscope mounted on the skin above the upper sternum was proposed to recognize fall-like activities of daily living (ADL) events. The output signals from the tri-axial accelerometer and bi-axial gyroscope were obtained during eight falls and eleven ADL action sequences. The thresholds of signal vector magnitude (SVM_Acc), angular velocity (${\omega}_{res}$), and angular variation (${\theta}_{res}$) were calculated using MATLAB. When the measured values of SVM_Acc, ${\omega}_{res}$, and ${\theta}_{res}$ were compared to the threshold values (TH1, TH2, and TH3), fall-like ADL events could be distinguished from a fall. When SVM_Acc was larger than 2.5 g (TH1), ${\omega}_{res}$ was larger than 1.75 rad/s (TH2), and ${\theta}_{res}$ was larger than 0.385 rad (TH3), eight falls and eleven ADL action sequences were recognized as falls. When at least one of these three conditions was not satisfied, the action sequences were recognized as ADL. Fall-like ADL events such as jogging and jumping up (or down) have posed a problem in distinguishing ADL events from an actual fall. When the measured values of SVM_Acc, ${\omega}_{res}$, and ${\theta}_{res}$ were applied to the sequential processing algorithm proposed in this study, the sensitivity was determined to be 100% for the eight fall action sequences and the specificity was determined to be 100% for the eleven ADL action sequences.

• 한국응급구조학회지
• /
• 제20권3호
• /
• pp.69-83
• /
• 2016

Purpose: The purpose of this study was to explore the factors influencing evidence-based fall prevention nursing performance of hospital nurses. Methods: A self-reported questionnaire was completed by 344 nurses from three general hospitals from January 20 to March 10, 2013. The study instruments included general characteristics of the subjects, and awareness and performance of fall prevention. Data were analyzed by t test, ANOVA, Pearson's correlation, and multiple regression using SPSS v. 20.0. Results: There were statistically significant differences in awareness and performance according to age, marital status, clinical experiences, workplace, experience of fall prevention education, knowledge of fall prevention, compliance with fall prevention, attention level toward prevention, recognition level of potential falls, nurse responsibility for falls, importance of fall prevention, efforts level for fall prevention, and awareness score of falls prevention. There was a positive correlation among awareness and performance of fall prevention. Based on the multiple regression analysis, compliance with fall prevention, efforts level for fall prevention, and awareness score of falls prevention were significant predictors for performance of fall prevention. The explanation power of the model was 64.1%. Conclusion: The findings revealed the need to develop an effective nursing intervention to improve hospital nurses' performance for fall prevention.

• 한국정밀공학회지
• /
• 제19권1호
• /
• pp.158-164
• /
• 2002

This paper proposes the optimization of ultrasonic signal recognition in ferrite carbon steel weld zone using neural networks. For these purposes, the ultrasonic signals for defects as porosity, incomplete penetration and slag inclusion in the weld zone are acquired in the type of time series data. And then their applications evaluated feature extraction based on the time-frequency-attractor domain(peak to peak, rise time, rise slope, fall time, fall slope, pulse duration, power spectrum, and bandwidth) and attractor characteristics (fractal dimension and attractor quadrant) etc. The proposed neural networks system in this study can enhances performance of ultrasonic signal recognition.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제16권12호
• /
• pp.3991-4007
• /
• 2022

In the public healthcare, a computational system that can automatically and efficiently detect and classify falls from a video sequence has significant potential. With the advancement of deep learning, which can extract temporal and spatial information, has become more widespread. However, traditional 3D CNNs that usually adopt shallow networks cannot obtain higher recognition accuracy than deeper networks. Additionally, some experiences of neural network show that the problem of gradient explosions occurs with increasing the network layers. As a result, an enhanced three-dimensional ResNet-based method for fall detection (3D-ERes-FD) is proposed to directly extract spatio-temporal features to address these issues. In our method, a 50-layer 3D residual network is used to deepen the network for improving fall recognition accuracy. Furthermore, enhanced residual units with four convolutional layers are developed to efficiently reduce the number of parameters and increase the depth of the network. According to the experimental results, the proposed method outperformed several state-of-the-art methods.

• 근관절건강학회지
• /
• 제26권1호
• /
• pp.54-61
• /
• 2019

Purpose: This study was performed to identify nursing students' fall prevention activities, health beliefs of falling and factors associated with fall prevention activities among nursing students. Methods: 149 nursing students from a university completed self-administered questionnaires including participants' characteristics, fall prevention activities, and health belief of falling. Hierarchical multiple regression analysis was used to determine significant independent factors of fall prevention activities. Results: The score for fall prevention activities was $62.40{\pm}9.78$, which was relatively high. The regression model had an adjusted $R^2$ of .16, which indicated that perceived susceptibility was a factor affecting fall prevention activities of nursing students. Conclusion: To increase perceived susceptibility, repetitive fall prevention education including various examples of falls could help nursing students to promote fall prevention activities. Nursing faculty should develop contents to increase recognition of obligation and responsibility regarding fall prevention in curriculum for nursing students.