• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• fall
• /
• pp.208-211
• /
• 2021

본 논문에서는 합성곱 신경망을 활용하여 영상에서 마스크 착용 및 미착용 상태를 탐지하는 방법을 제안한다. 코로나바이러스감염증-19(COVID-19)의 유행에 따라 감염 및 확산방지를 위해 마스크 정상적 착용이 요구되는데 몇몇 사람들은 이를 지키지 않고 있으며 현재의 감시 시스템은 입구에서 마스크 착용 여부를 검사하는 방식으로 작동될 뿐 공간에 입장한 다음 착용 여부를 알 수 없다. 제안하는 방법은 합성곱 신경망을 통해 영상에서 얼굴을 탐지하여 얻은 데이터를 이용하여 다수사람들의 마스크 착용 및 미착용 상태를 판별하는 방법으로 설계하였다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.3
• /
• pp.333-340
• /
• 2022

The COVID-19 has made everyone anxious and people need to keep their distance. It is necessary to conduct collective assessment and screening of college students' mental health in the opening season of every year. This study uses and trains a multi-layer perceptron neural network model for deep learning to identify facial emotions. After the training, real pictures and videos were input for face detection. After detecting the positions of faces in the samples, emotions were classified, and the predicted emotional results of the samples were sent back and displayed on the pictures. The results show that the accuracy is 93.2% in the test set and 95.57% in practice. The recognition rate of Anger is 95%, Disgust is 97%, Happiness is 96%, Fear is 96%, Sadness is 97%, Surprise is 95%, Neutral is 93%, such efficient emotion recognition can provide objective data support for capturing negative. Deep learning emotion recognition system can cooperate with traditional psychological activities to provide more dimensions of psychological indicators for health.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• fall
• /
• pp.290-293
• /
• 2021

In this paper, we implement a learning efficiency verification system to inspire learning motivation and help improve concentration by detecting the situation of the user studying. To this aim, data on learning attitude and concentration are measured by extracting the movement of the user's face or body through a real-time camera. The Jetson board was used to implement the real-time embedded system, and a convolutional neural network (CNN) was implemented for image recognition. After detecting the feature part of the object using a CNN, motion detection is performed. The captured image is shown in a GUI written in PYQT5, and data is collected by sending push messages when each of the actions is obstructed. In addition, each function can be executed on the main screen made with the GUI, and functions such as a statistical graph that calculates the collected data, To do list, and white noise are performed. Through learning efficiency checking system, various functions including data collection and analysis of targets were provided to users.

• Journal of Korea Multimedia Society
• /
• v.19 no.11
• /
• pp.1852-1861
• /
• 2016

Drowsy driving is a significant factor in traffic accidents, so driver drowsiness detection system based on computer vision for convenience and safety has been actively studied. However, it is difficult to accurately detect the driver drowsiness in complex background and environmental change. In this paper, it proposed the driver drowsiness detection algorithm to determine whether the driver is drowsy through the measurement standard of a yawn, eyes drowsy status, and nod based on facial features. The proposed algorithm detect the driver drowsiness in the complex background, and it is robust to changes in the environment. The algorithm can be applied in real time because of the processing speed faster. Throughout the experiment, we confirmed that the algorithm reliably detected driver drowsiness. The processing speed of the proposed algorithm is about 0.084ms. Also, the proposed algorithm can achieve an average detection rate of 98.48% and 97.37% for a yawn, drowsy eyes, and nod in the daytime and nighttime.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.18 no.6
• /
• pp.731-736
• /
• 2008

In this paper, we propose the movement detection algorithm by using virtual skeleton model. To do this, first, we eliminate error values by using conventioanl method based on RGB color model and eliminate unnecessary values by using the HSI color model. Second, we construct the virtual skeleton model with skeleton information of 10 peoples. After matching this virtual model to original image, we extract the real head silhouette by using the proposed circle searching method. Third, we extract the object by using the mean-shift algorithm and this head information. Finally, we validate the applicability of the proposed method through the various experiments in a complex environments.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.19 no.5
• /
• pp.609-614
• /
• 2009

The purpose of this paper is to design the metal object detection system for drive inside protection. To do this, we propose the algorithm for designing the color filter that can detect the metal object using fuzzy theory and the algorithm for detecting area of the driver's face using fuzzy skin color filter. Also, by using the proposed algorithm, we propose the algorithm for detecting the metallic object candidate regions. And, the metallic object color filter is then applied to find the candidate regions. Finally, we show the effectiveness and feasibility of the proposed method through some experiments.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.540-542
• /
• 2023

Due to the serious issues posed by facial manipulation technologies, many researchers are becoming increasingly interested in the identification of face forgeries. The majority of existing face forgery detection methods leverage powerful data adaptation ability of neural network to derive distinguishing traits. These deep learning-based detection methods frequently treat the detection of fake faces as a binary classification problem and employ softmax loss to track CNN network training. However, acquired traits observed by softmax loss are insufficient for discriminating. To get over these limitations, in this study, we introduce a novel discriminative feature learning based on Vision Transformer architecture. Additionally, a separation-center loss is created to simply compress intra-class variation of original faces while enhancing inter-class differences in the embedding space.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.18 no.5
• /
• pp.1171-1176
• /
• 2014

The eye-tracking [ET] is used on the human computer interaction [HCI] analysing the movement status as well as finding the gaze direction of the eye by tracking pupil's movement on a human face. Nowadays, the ET is widely used not only in market analysis by taking advantage of pupil tracking, but also in grasping intention, and there have been lots of researches on the ET. Although the vision based ET is known as convenient in application point of view, however, not robust in changing environment such as illumination, geometrical rotation, occlusion and scale changes. This paper proposes two steps in the ET, at first, face and eye regions are discriminated by Haar classifier on the face, and then the pupils from the discriminated eye regions are tracked by CAMShift as well as Template matching. We proved the usefulness of the proposed algorithm by lots of real experiments in changing environment such as illumination as well as rotation and scale changes.

• The KIPS Transactions:PartA
• /
• v.11A no.7 s.91
• /
• pp.529-536
• /
• 2004

Smart homes integrated with sensors, actuators, wireless networks and context-aware middleware will soon become part of our daily life. This paper describes a context-aware middleware providing an automatic home service based on a user's preference. The context-aware middle-ware utilizes 6 basic data for learning and predicting the user's preference on the multimedia content : the pulse, the body temperature, the facial expression, the room temperature, the time, and the location. The six data sets construct the context model and are used by the context manager module. The log manager module maintains history information for multimedia content chosen by the user. The user-pattern learning and pre-dicting module based on a neural network predicts the proper home service for the user. The testing results show that the pattern of an in-dividual's preferences can be effectively evaluated and predicted by adopting the proposed context model.

• Journal of Satellite, Information and Communications
• /
• v.11 no.3
• /
• pp.18-23
• /
• 2016

Internet of things (IoT) is revolutionizing in the patient-Centered medical monitoring and management by authorizing the Smartphone application and data analysis with medical centers. The network connectivity is basic requirement to collect the observed human beings' health information from Smartphone to monitor the health from IoT medical devices in personal healthcare. The IoT environment built in Smartphone is very effective and does not demand infrastructure. This paper presents the smart phone deployed personal IoT architecture for Non-Invasive ECG Capturing. The adaptable IoT medical device cum Gateway is used for personal healthcare with big data storage on cloud configuration. In this approach, the Smartphone camera based imaging technique used to extract the personal ECG waveform and forward it to the cloud based big data storage connectivity using IoT architecture. Elaborated algorithm allows for efficient ECG registration directly from face image captured from Smartphone or Tablet camera. The profound technique may have an exceptional value in monitoring personal healthcare after adequate enhancements are introduced.