• Journal of the Korea Society of Computer and Information
• /
• v.26 no.5
• /
• pp.1-8
• /
• 2021

Coronavirus disease 2019 (COVID-19) has affected the world seriously. Every person is required for wearing a mask properly in a public area to prevent spreading the virus. However, many people are not wearing a mask properly. In this paper, we propose an efficient mask detection system. In our proposed system, we first detect the faces of input images using YOLOv5 and classify them as the one of three scene complexity classes (Simple, Moderate, and Complex) based on the number of detected faces. After that, the image is fed into the Faster-RCNN with the one of three ResNet (ResNet-18, 50, and 101) as backbone network depending on the scene complexity for detecting the face area and identifying whether the person is wearing the mask properly or not. We evaluated our proposed system using public mask detection datasets. The results show that our proposed system outperforms other models.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.17 no.7
• /
• pp.492-498
• /
• 2016

It is well known that since abrupt faults in induction motors tend to lead to subsequent faults and deterioration of the drive apparatus, motor faults may lead to several operating restrictions, such as security problems and economic loss. A lot of research has been done in the area of diagnosis to detect machine faults and to prevent catastrophic hazards in the motor drive system. This paper presents a new method of motor current signature analysis in which the DC-link current of the inverter-driven induction motor system, where a single current sensor is employed instead of three AC current sensors, is measured, and fast Fourier transform analysis is performed. This proposed method makes it possible to easily discern and clearly separate the motor fault current signature from the normal operation current flowing through the stator and rotor windings.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.47 no.5
• /
• pp.16-23
• /
• 2010

This paper presents 6bit 250MS/s flash A/D converter which can be applied to wireless communication system. To solve the problem of large power consumption in flash A/D converter, control algorithm by input signal level is used in comparator stage. Also, input voltage range detector circuit is used in reference resistor array to minimize the dynamic power consumption in the comparator. Compared with the conventional A/D converter, the proposed A/D converter shows 4.3% increase of power consumption in analog and a seventh power consumption in digital, which leads to a half of power consumption in total. The A/D converter is implemented in a $0.18{\mu}m$ CMOS 1-poly 6-metal technology. The measured results show 106mW power dissipation with 1.8V supply voltage. It shows 4.1bit ENOB at sampling frequency 250MHz and 30.27MHz input frequency.

• Journal of Digital Contents Society
• /
• v.3 no.2
• /
• pp.255-264
• /
• 2002

Multimedia is now applied to various real worlds. In particular, the focus of CSCW(Computer Supported Cooperated Work) for multimedia education system has increased. DOORAE is a framework for supporting development of applications running on distributed multimedia environment and multimedia distance education system. EDA is a system is able to detect automatically a software error based on distributed multimedia. It has been designed and implemented for construction and experiment of effective DOORAE environment. It detects an error by polling periodically the process with relation to session. Conventional method detects an error by polling periodically all the process. This papaer explains a performance analysis of an error detection system running on distributed multimedia environment using the rule-based SES and DEVS modeling and simulation techniques. In DEVS, a system has a time base, inputs, states, outputs, and functions.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.37 no.3
• /
• pp.57-69
• /
• 2000

A moving window technique for detecting a lane and obstacles using the Images captured by a CCD camera attached in an automobile, is proposed in this paper To process the dynamic images in real time, there could be many constraints on the hardware To overcome these hardware constraints and to detect the lane and obstacles in real time, the optimal size of window IS determined based upon road conditions and automobile states. By utilizing the sub-Images inside the windows, detection of the lane and obstacles become possible m real time. For each Image frame, the moving windows are re-determined following the predicted directions based on Kalman filtering theory to Improve detection accuracy, as well as efficiency The feasibility of proposed algorithm IS demonstrated through the simulated experiments of highway driving.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.37 no.6
• /
• pp.551-555
• /
• 2009

We studied the design of fiber optic gyroscope that enables to sense high rotation by extending the limit of rotation sensibility of fiber optic interferometer. Based on the digital serrodyne modulation technique, the signal processing of fiber optic gyroscope was designed and the prototype fiber optic gyroscope showed the high rotation sensibility up to ${\pm}3000[deg/sec]$ and scale factor performance of about 150[ppm] by the experiments. Accordingly, we confirmed that the design of fiber optic gyroscope was valid for high rotation.

• Journal of Satellite, Information and Communications
• /
• v.12 no.2
• /
• pp.56-60
• /
• 2017

This paper describes the design and implementation of fall-down detection algorithm based on image processing. The fall-down detection algorithm separates objects by using background subtraction and binarization after grayscale conversion of the input image acquired by the camera, and recognizes the human body by using labeling operation. The recognized human body can be monitored on the display image, and an alarm is generated when fall-down is detected. By using computer simulation, the proposed algorithm has shown a detection rate of 90%. We verify the feasibility of the proposed system by verifying the function by using the prototype test implemented on the DSP image processing board.

• Journal of IKEEE
• /
• v.27 no.3
• /
• pp.308-312
• /
• 2023

The gate current sensing structure was proposed to more effectively control the regulation of the output voltage when the LDO regulator occurs in an overshoot or undershoot situation. In a typical existing LDO regulator, the regulation voltage changes when the load current changes. However, the operation speed of the pass transistor can be further improved by supplying/discharging the gate terminal current in the pass transistor using a gate current sensing structure. The input voltage of the LDO regulator using the gate current sensing structure is 3.3 V to 4.5 V, the output voltage is 3 V, and the load current has a maximum value of 250 mA. As a result of the simulation, a voltage change value of about 12 mV was confirmed when the load current changed up to 250 mA.

• Proceedings of the Korean Information Science Society Conference
• /
• 2004.04b
• /
• pp.772-774
• /
• 2004

선택적 주의선택 알고리즘 중 색상과 밝기 지도를 만드는 부분을 실시간 처리에 적용하고, 저주파수가 주로 분포하는 영역과 물체의 움직임이 감지된 영역을 입력 영상에서 찾아내어 가중치별로 합산함으로서 실시간으로 선택적 주의를 줄 수 있는 시스템을 구현했다.

• 한국HCI학회:학술대회논문집
• /
• 2009.02a
• /
• pp.62-67
• /
• 2009

Many interaction methods for augmented reality has attempted to reduce difficulties in tracking of interaction subjects by either allowing a limited set of three dimensional input or relying on auxiliary devices such as data gloves and paddles with fiducial markers. We propose Spatial Interaction (SPINT), a noncontact passive method that observes an occupancy state of the spaces around target virtual objects for interpreting user input. A depth-sensing camera is introduced for constructing the virtual space sensors, and then manipulating the augmented space for interaction. The proposed method does not require any wearable device for tracking user input, and allow versatile interaction types. The depth perception anomaly caused by an incorrect occlusion between real and virtual objects is also minimized for more precise interaction. The exhibits of dynamic contents such as Miniature AR System (MINARS) could benefit from this fluid 3D user interface.