• 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.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.11a
• /
• pp.177-180
• /
• 2021

코로나바이러스로 인하여 전 세계는 어려움을 겪고 있으며, 바이러스를 확산을 막기 위해서 실외에서는 마스크를 쓰는 것이 일상이 되었다. 하지만, 이를 따르지 않는 사람이 일반 시설에 방문할 때 이를 감지하고 경고를 할 수 있는 시스템이 없어, 마스크 미 착용자로 인한 위험성 방지에 취약점을 가지고 있다. 본 연구에서는 OpenCV 라이브러리를 이용한 마스크 착용 여부를 확인하는 시스템을 설계한다. Haar 특징기반 다단계 분류자를 이용하여 마스크 인식 프로세스를 설계하였으며, 마스크 착용 확인 시스템은 경량 컴퓨터인 라즈베리파이 장치 위에 구현하였다. 또한 확인된 사람의 이미지는 클라우드 시스템에 저장할 수 있도록 구현하였다. 본 연구를 통해, 누구나 손쉽게 해당 마스크 착용 확인 시스템을 중소 매장에 설치하여 사용할 수 있으며, 코로나바이러스 확산 방지에 기여할 수 있다고 예상한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2007.11a
• /
• pp.80-82
• /
• 2007

외곽 검출을 통한 스테레오 비전의 실시간 거리 측정을 제안하였다. 주차 시스템이나 이동 로봇의 경로 설정에 있어 실시간으로 거리 측정이 중요시 되고, 이를 해결하기 위하여 인간의 시각에 가까운 평행 축 상의 두 대의 카메라에서 취득된 영상을 이용한다. 카메라에서 얻은 영상으로부터 외곽을 추출하기 위해 Sobel Mask를 사용하였으며, 좌 우 구분을 위해 색 변환과 영상 정보로부터 거리를 측정하기 위해 영상 등록 과정을 거치고 거리 측정을 하였다. 거리 측정의 결과 2.4%의 오차율을 보였으며 이는 로봇의 이동 간에 적용할 경우 양호한 결과를 얻을 수 있을 것이다.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.42 no.3 s.303
• /
• pp.91-98
• /
• 2005

In this paper, we propose a temporal error concealment method based on the edge adaptive masking. In the method, four regions around the corrupted block - top, bottom, left, and right - are defined and the edge features of the regions are extracted by applying an edge operator for each direction. The size of a mask for the boundary matching is determined by the edge information, which can be considered as a criterion to measure the activity of the boundary region. In other words, it is determined such that the size of the mask is proportional to the amount of edge-component extracted from each region in order to yield the higher reliability on boundary matching. This process is equivalent to applying weights depending on the edge features, which leads the improved motion vector. In experiments, it is verified that the proposed method outperforms the conventional methods in terms of image quality, and then its merits and demerits are discussed.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.32 no.11A
• /
• pp.1086-1095
• /
• 2007

This paper proposes a novel non-coherent TOA estimation scheme using multiple correlation process on the existence of multiple simultaneously operating piconets (SOPs). Impulse radio-ultra wideband (IR-UWB) based on direct sequence spread spectrum (DSSS) using Gold sequence is employed in order to discriminate each piconet. In order to enhance the characteristic of correlation, this paper presents the method of multiple mask operation (MMO). The time of arrival (TOA) of direct line of sight (DLOS) path is estimated via two step coarse/fine timing detection. To verify the performance of proposed scheme, two distinct channel models approved by IEEE 802.15.4a Task Group (TG) are considered. According to the simulation results, it could conclude that the proposed scheme have performed better performance than the conventional method well even in densed indoor multi-path environment as well as in the existence of multiple SOPs.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.24 no.4
• /
• pp.327-334
• /
• 2006

The aerial photos need in a simple object such as cartography and ground cover classification and also in a social objects such as the city plan, environment, disaster, transportation etc. However, the shadow, which includes when taking the aerial photos, makes a trouble to interpret the ground information, and also users, who need the photos in their field tasks, have a restriction. Generally the shadow occurs by the building and surface topography, and the detail cause is by changing of the illumination in an area. For removing the shadow this study uses the single image and processes the image without the source of image and taking situation. Also, applying the entropy minimization method it generates the 1-D gray-scale invariant image for creating the shadow edge mask and using the Canny edge detection creates the shadow edge mask, and finally by filtering in Fourier frequency domain creates the intrinsic image which recovers the 3-D color information and removes the shadow.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.50 no.5
• /
• pp.231-241
• /
• 2013

In this paper, we propose an average approximation filter which can effectively remove the noises of the images. The noises include impulse noises, gaussian noises and mixed noises. The algorithm is as follows. First, as a step of noise detection, we find whether the difference between the pixel value and the average value is greater than the threshold value or not after getting the average value that removed the minimum and maximum values in the applied mask. If the pixel value is greater than the threshold value, the pixel value is processed as noise. If it is less than or equal to the threshold value, it is processed as non-noise. Next, as the noise reduction step, we output the approximate value in mask as the pixel value and the average value except the minimum and maximum values of the pixel including the noise. As the result of applying this average approximation filter to the mixed noise images, the approximation filter can reduce the noises effectively more than 0.4[dB] as compared with applying the median filter and the average filter, respectively.

• Journal of the Korea Society of Computer and Information
• /
• v.12 no.5
• /
• pp.313-319
• /
• 2007

In this paper, we propose a novel method to extract an area of car licence plate and codes of vehicle number from a photographed car image using features on vertical edges and a new Fuzzy neural network algorithm to recognize extracted codes. Prewitt mask is used in searching for vertical edges for detection of an area of vehicle number plate and feature information of vehicle number palate is used to eliminate image noises and extract the plate area and individual codes of vehicle number. Finally, for recognition of extracted codes, we use the proposed Fuzzy neural network algorithm, in which FCM is used as the learning structure between input and middle layers and Max_Min neural network is used as the learning structure within inhibition and output layers. Through a variety of experiments using real 150 images of vehicle, we showed that the proposed method is more efficient than others.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.53 no.2
• /
• pp.36-44
• /
• 2016

In this paper, an Impulse Radio-Ultra Wide band RF Transceiver for WBAN applications is implemented in $0.18{\mu}m$ CMOS technology. The designed RF transceiver support 3-5GHz UWB low band and employs OOK(On-Off Keying) modulation. The receiver employs non-coherent energy detection architecture to reduce complexity and power consumption. For the rejection of the undesired interferers and improvement of the receiver sensitivity, RF active notch filter is integrated. The VCO based transmitter employs the switch mechanism. As adapt the switch mechanism, power consumption and VCO leakage can be reduced. Also, the spectrum mask is always same at each center frequency. The measured sensitivity of the receiver is -84.1 dBm at 3.5 GHz with 1.579 Mbps. The power consumption of the transmitter and receiver are 0.3nJ/bit and 41 mW respectively.

• Journal of Convergence for Information Technology
• /
• v.11 no.6
• /
• pp.213-218
• /
• 2021

In this paper, we detect and track vehicles by type based on highway daytime driving videos taken with black boxes for vehicles. In addition, we design a real-time MR contents production method that can be newly created by placing substitute videos of each type of detected vehicles in the same location as the new background video. To detect and track vehicles by type, we use the YOLO algorithm. And we also use the mask technique based on RGB color for substitute videos of each type of vehicles detected. The size of the vehicle substitute videos to be used for MR content are substituted by the same size as the area size of the detected vehicles. In this paper, we confirm that real-time MR contents design is possible as a result of experiments and simulations and believe that It will be usefully utilized in the field of VR contents.