• 한국지능시스템학회논문지
• /
• 제26권1호
• /
• pp.70-75
• /
• 2016

머신비전을 사용하여 사람의 얼굴을 인식하는 다양한 연구가 진행되고 있다. 머신비전은 기계에 시각을 부여하여 이미지를 분류 혹은 분석하는 기술을 의미한다. 본 논문에서는 이러한 머신비전 기술을 적용한 얼굴을 분류하는 알고리즘을 제안한다. 이 얼굴 분류 알고리즘을 구현하기 위해 컨볼루셔널 신경망(Convolution neural network)과 Cascade 안면 검출기를 사용하였고, 피험자들의 얼굴을 분류하였다. 구현한 얼굴 분류 알고리즘의 학습을 위해 한 피험자 당 이미지 2,000장, 3,000장, 40,00장을 10회와 20회 컨볼루셔널 신경망에 각각 반복하여 학습과 분류를 진행하였고, 학습된 컨볼루셔널 신경망과 얼굴 분류 알고리즘의 실효성을 테스트하기 위해 약 6,000장의 이미지를 분류하였다. 또한 USB 카메라 영상을 실험 데이터로 입력받아 실시간으로 얼굴을 검출하고 분류하는 시스템을 구현하였다.

• Journal of the Korean Academy of Child and Adolescent Psychiatry
• /
• 제25권3호
• /
• pp.113-120
• /
• 2014

This is an analysis report of the "Korean Academy of Child and Adolescent Psychiatry (KACAP), Vision 2033 Survey". The survey questionnaires were developed by the planning department of KACAP and sent to KACAP members from 2012 to 2013. This survey consisted of six categories : membership, academic activity, journal publication, administrative system, fellowship training program, and future planning. The response rate was 40.5%. In addition to multiple choice questions, responders also described their own ideas and suggestions regarding KACAP. The results of this study can be used as evidence for planning the vision 2033 of KACAP.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1996년도 하계학술대회 논문집 B
• /
• pp.1119-1121
• /
• 1996

The main purpose of this study is to achieve the effective port works by using of container-crane, to disposer of many containers rapidly by using of vision sensor in order to control the swing of spreader. It is examined the possibility of automation in container-crane through a test in the field.

• 반도체디스플레이기술학회지
• /
• 제20권3호
• /
• pp.120-124
• /
• 2021

Recently, human attribute recognition (HAR) attracts a lot of attention due to its wide application in video surveillance systems. Recent deep-learning-based solutions for HAR require time-consuming training processes. In this paper, we propose a post-processing technique that utilizes the historical video frames to improve prediction results without invoking re-training or modifying existing deep-learning-based classifiers. Experiment results on a large-scale benchmark dataset show the effectiveness of our proposed method.

• 한국안광학회지
• /
• 제20권4호
• /
• pp.491-499
• /
• 2015

목적: 시각이상을 가진 아동을 대상으로 시기능훈련을 통한 시각기능이상의 개선 전후 시지각점수의 변화를 살펴보고 시지각에 미치는 영향을 알아보고자 한다. 방법: 시각기능이상을 가진 13세 미만($8.75{\pm}1.66$)의 아동 23명을 대상으로 시기능훈련 전후에 시지각기능검사(TVPS_R; test of visual perceptual skills-revised)를 실시하여 시기능훈련에 따른 시각기능의 변화와 시지각 평가점수의 변화를 분석하였다. 결과: 시기능훈련을 통해서 원거리 PRC(positive relative convergence) 분리점은 평균 $3.39{\pm}2.59{\Delta}$(prism)에서 $13.87{\pm}6.04{\Delta}$로 증가되었고, 근거리 P.R.C 분리점은 평균 $5.48{\pm}3.42{\Delta}$에서 $18.44{\pm}7.58{\Delta}$로 증가하였다. NPC(near point of convergence)는 $25.87{\pm}7.33cm$에서 $7.48{\pm}2.83cm$로 개선되었고, NPA(near point of accommodation)는 $19.57{\pm}7.16cm$에서 $7.09{\pm}1.88cm$로 개선되었다. 시지각평가에서 시각기억분야를 제외하고는 대응차가 시각완성에서 $17.74{\pm}16.94$(p=0.000), 시각적순차기억에서 $15.65{\pm}17.11$(p=0.000), 배경식별에서 $13.65{\pm}16.63$(p=0.001), 형태향상성에서 $12.74{\pm}18.41$(p=0.003), 시각구별에서 $6.49{\pm}10.07$(p=0.005), 시각적공간지각에서 $4.17{\pm}9.33$(p=0.043) 순으로 개선되었고, 이를 종합한 시지각점수는 대응차가 $15.22{\pm}8.66$(p=0.000)로서 더욱 더 유의한 결과를 나타내었다. 결론: 시기능훈련을 통해서 시각기능의 개선과 시각기억분야를 제외한 시지각점수의 향상이 시각완성, 시각적순차기억, 배경식별, 형태향상성, 시각구별, 시각적공간지각 순으로 유의하게 나타났다. 따라서 시각기능향상을 위한 시기능훈련은 시각의 기능뿐만 아니라 시지각의 기능까지 영향을 미치는 것을 확인 할 수 있으며 시각훈련의 중요성을 인식 할 수 있었다.

• 한국안광학회지
• /
• 제19권2호
• /
• pp.247-252
• /
• 2014

목적: 단기간의 중심외주시 훈련 후 발생하는 신경의 재조직화의 특성과 호발 위치를 알아보고자 했다. 방법: 정상시력을 가진 성인 14명을 대상으로 21일 간 중심외주시 훈련을 하고 훈련 전후 광지각도와 다국소망막전위도를 측정하여 사후 분석했다. 결과: 중심외주시 훈련 전후 값을 비교한 광지각도 검사(p<0.047)에서 뿐만 아니라 다국소망막전위도 검사에서도 유의한 개선을 보였다(p<0.028). 결론: 시각 말초신경계는 재생이 불가능하지만 단기간의 중심외주시 훈련은 말초신경계에서 신경 재조직화를 발생시킨다.

• International journal of advanced smart convergence
• /
• 제6권1호
• /
• pp.18-25
• /
• 2017

Recently, VR HMD (virtual reality head mounted display) has been utilized for virtual training, entertainment, vision therapy, and optometry. In particular, virtual reality contents are increasingly used for vision therapy and optometry. Accordingly, high-quality virtual reality contents such as a natural vision of life is required. Therefore, it is necessary to study the content production according to the optical characteristics of the VR HMD. The purpose of this paper is to suggest a proper FOV (field of view) of contents according to the distortion rate. We produced virtual reality contents and obtained distorted images by virtual camera. The distortion rate is calculated by using the distorted image. It is proved that the optimal FOV of the VR content with the minimum distortion is $90{\sim}100^{\circ}$. The results of this study are expected to be applied to the production of high quality contents.

• 반도체디스플레이기술학회지
• /
• 제21권1호
• /
• pp.75-78
• /
• 2022

Deep Neural Networks (DNN) has become an essential data processing architecture for the implementation of multiple computer vision tasks. Recently, DNN-based algorithms achieve much higher recognition accuracy than traditional algorithms based on shallow learning. However, training and inference DNNs require huge computational capabilities than daily usage purposes of computers. Moreover, with increased size and depth of DNNs, CPUs may be unsatisfactory since they use serial processing by default. GPUs are the solution that come up with greater speed compared to CPUs because of their Parallel Processing/Computation nature. In this paper, we analyze the inference time complexity of DNNs using well-known computer vision library, OpenCV. We measure and analyze inference time complexity for three cases, CPU, GPU-Float32, and GPU-Float16.

• 국제학술발표논문집
• /
• The 7th International Conference on Construction Engineering and Project Management Summit Forum on Sustainable Construction and Management
• /
• pp.252-259
• /
• 2017

Nowadays, the existing manual method for recording actual progress of the construction site has some drawbacks, such as great reliance on the experience of professional engineers, work-intensive, time consuming and error prone. A method integrating computer vision and BIM(Building Information Modeling) is presented for indoor automatic progress monitoring. The developed method can accurately calculate the engineering quantity of target component in the time-lapse images. Firstly, sample images of on-site target are collected for training the classifier. After the construction images are identified by edge detection and classifier, a voting algorithm based on mathematical geometry and vector operation will divide the target contour. Then, according to the camera calibration principle, the image pixel coordinates are conversed into the real world Coordinate and the real coordinates would be corrected with the help of the geometric information in BIM model. Finally, the actual engineering quantity is calculated.

• International Journal of Computer Science & Network Security
• /
• 제24권6호
• /
• pp.89-98
• /
• 2024

Due to the increase in population and consequently the increase in the production of plastic waste, recovery of this part of the waste is an undeniable necessity. On the other hand, the recycling of plastic waste, if it is placed in a systematic process and controlled, can be effective in creating jobs and maintaining environmental health. Waste collection in many large cities has become a major problem due to lack of proper planning with increasing waste from population accumulation and changing consumption patterns. Today, waste management is no longer limited to waste collection, but waste collection is one of the important areas of its management, i.e. training, segregation, collection, recycling and processing. In this study, a systematic method based on machine vision for sorting plastic bottles in different colors for recycling purposes will be proposed. In this method, image classification and segmentation techniques were presented to improve the performance of plastic bottle classification. Evaluation of the proposed method and comparison with previous works showed the proper performance of this method.