• Proceedings of the Korea Information Processing Society Conference
• /
• 2018.10a
• /
• pp.736-738
• /
• 2018

최근 딥러닝의 발전에 따라 무인감시, CCTV 등 영상감시 시스템도 지능화되고 있다. 하지만 쓰레기 무단투기 감시는 여전히 관리자가 실시간으로 CCTV 영상을 관제하는 형태로 이루어지고 있다. 이러한 문제를 해결하기 위해 본 논문에서는 CCTV 영상에서 쓰레기 무단투기를 검출하는 방법을 제안하며 검출 방법으로 차 영상과 합성곱 신경망을 이용한다. 실험은 합성곱 신경망에서의 쓰레기봉투 분류 문제 위주로 진행하였다. 합성곱 신경망의 네트워크는 Inception v3를 사용하였으며 실험 결과, 약 99.52%의 쓰레기봉투 분류율을 얻을 수 있었다.

• Proceedings of the Korea Multimedia Society Conference
• /
• 2003.05b
• /
• pp.431-434
• /
• 2003

In this paper, we propose an efficient method not only f3r producing accurate region segmentation, solving the over-segmentation problem of watershed algorithm but also f3r reducing post-processing time by reducing computation loads. Through this proposed method, region segmentation of neighboring objects and discrimination of similar intensities were effectively obtained. Input image of watershed algorithm has used the derivative-based detectors such as Sobel and Canny. But proposed method uses the pixels-similarity-based detector, that is, SUSAN. By adopting this proposed method, we can reduce the noise problem and solve the problem of over-segmentation and not lose the edge information of objects. We also propose Zero-Crossing SUSAN. With Zero-Crossing SUSAN, the edge localization, times and computation loads can be improved over those obtained from existing SUSAN

• Journal of Korea Multimedia Society
• /
• v.11 no.12
• /
• pp.1649-1657
• /
• 2008

In this paper, to overcome the failure of binarization for the characters suffered from low contrast and non-uniform illumination in license plate character recognition system, we improved the binarization method by combining local thresholding with global thresholding and edge detection. Firstly, apply the local thresholding method to locate the characters in the license plate image and then get the threshold value for the character based on edge detector. This method solves the problem of local low contrast and non-uniform illumination. Finally, back-propagation Neural Network is selected as a powerful tool to perform the recognition process. The results of the experiments i1lustrate that the proposed binarization method works well and the selected classifier saves the processing time. Besides, the character recognition system performed better recognition accuracy 95.7%, and the recognition speed is controlled within 0.3 seconds.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2015.04a
• /
• pp.870-872
• /
• 2015

본 연구에서는 웨이퍼의 자동광학검사를 위한 결점 검출 비전 시스템을 개발하였다. 성공적 결점 검출을 위해 몇 가지 이진화 방법을 비교하였고, bimodal과 unimodal 분포에 모두 좋은 결과를 나타낸 개선된 Otsu 방법을 선택하였다. 빠르고 정확한 임계값 계산을 위해 ROI 추출기능을 개발하였으며 최종적으로 웨이퍼의 검출 패턴은 정의된 기준에 따라 영상 분류되었고 성능평가를 위해 14개 이상의 웨이퍼 영상으로 테스트하였다.

• Journal of Korea Multimedia Society
• /
• v.22 no.4
• /
• pp.424-431
• /
• 2019

In this paper, we propose a real-time hand gesture recognition algorithm to eliminate the inconvenience of using hand controllers in VR applications. The user's 3D hand coordinate information is detected by leap motion sensor and then the coordinates are generated into two dimensional image. We classify hand gestures in real-time by learning the imaged 3D hand coordinate information through SSD(Single Shot multibox Detector) model which is one of CNN(Convolutional Neural Networks) models. We propose to use all 3 channels rather than only one channel. A sliding window technique is also proposed to recognize the gesture in real time when the user actually makes a gesture. An experiment was conducted to measure the recognition rate and learning performance of the proposed model. Our proposed model showed 99.88% recognition accuracy and showed higher usability than the existing algorithm.

• Medical Lasers
• /
• v.9 no.1
• /
• pp.25-33
• /
• 2020

Optical-based imaging technology has high resolution and can assess images in real time. Numerous studies have been conducted for its application in the dental field. The current research introduces an oral camera that includes fluorescent imaging, a second study examining a 3D intraoral scanner applying a confocal method and a polarization structure that identifies the 3D image of a tooth, and finally, an optical coherence tomography technique. Using this technique, we introduce a new concept 3D oral scanner that simultaneously implements 3D structural imaging as well as images that diagnose the inside of teeth. With the development of light source technology and detector technology, various optical-based imaging technologies are expected to be applied in dentistry.

• Proceedings of the International Microelectronics And Packaging Society Conference
• /
• 2004.02a
• /
• pp.17-25
• /
• 2004

During the past few years, advances have been made in both in X-ray tube and detector technologies. The field of microfocus radioscopy has been established as an important testing process and has expanded into many new industrial applications that require quality control or process optimization. The first nanofocus and multifocus X-ray systems have become available with a focal spot of .5 micron. In the existing range of microfocus X-ray tubes, further improvements have been achieved as well, such as increased long term stability of intensity position constancy. Software, image processing and manipulation techniques have all progressed as well, allowing X-ray to become a formidable non-destructive inspection method for manufacturers in virtually every industry, especially those involved with Electronic Packaging and SMT.

• Journal of the Korean Society of Radiology
• /
• v.3 no.3
• /
• pp.37-40
• /
• 2009

In this study, we made a high efficiency x-ray detecting sensor using the lead oxide(PbO) that are used in direct method of x-ray detector. PbO with nano size particles is produced by sol-gel method for high efficiency. The produced PbO with nano size is deposited on ITO(Induim Tin Oxide) glass in several temperature using the PIB(particle-in-binder) method. The thickness of the deposited PbO is about $200{\mu}m$. Through the measurement of dark current, sensitivity and SNR(Signal To Noise Ratio), an electrical properties of the produced PbO film are analyzed. Therefore, we show that an electrical properties are changed according to a temperature and that the PbO film that was treated at $500^{\circ}C$ in O2 atmosphere is the most high efficiency x-ray detecting sensor.

• Journal of the Korean Society of Radiology
• /
• v.3 no.3
• /
• pp.27-35
• /
• 2009

In this study, the basic research verifying possibility of applications as radiology image sensor in Digital Radiography was performed, the radiology image sensor was fabricated using double layer technique tio decrease dark current. High efficiency material in substitution for a-Se have been studied as a direct method of imaging detector in Digital Radiography to decrease dark current by using Hetero junction already used as solar cell, semiconductor. Particle-In-Binder method is used to fabricate radiology image sensor because it has a lot of advantages such as fabrication convenient, high yield, suitability for large area sensor. But high leakage current is one of main problem in PIB method. To make up for the weak points, double layer technique is used, and it is considered that high efficient digital radiation sensor can be fabricated with easy and convenient process. In this study, electrical properties such as leakage current, sensitivity is measured to evaluate double layer radiation sensor material.

• Progress in Medical Physics
• /
• v.27 no.1
• /
• pp.1-7
• /
• 2016

In this work, we considered a compressed-sensing (CS)-based image deblurring scheme with a total-variation (TV) regularization penalty for improving image characteristics in digital tomosynthesis (DTS). We implemented the proposed image deblurring algorithm and performed a systematic simulation to demonstrate its viability. We also performed an experiment by using a table-top setup which consists of an x-ray tube operated at $90kV_p$, 6 mAs and a CMOS-type flat-panel detector having a $198-{\mu}m$ pixel resolution. In the both simulation and experiment, 51 projection images were taken with a tomographic angle range of ${\theta}=60^{\circ}$ and an angle step of ${\Delta}{\theta}=1.2^{\circ}$ and then deblurred by using the proposed deblurring algorithm before performing the common filtered-backprojection (FBP)-based DTS reconstruction. According to our results, the image sharpness of the recovered x-ray images and the reconstructed DTS images were significantly improved and the cross-plane spatial resolution in DTS was also improved by a factor of about 1.4. Thus the proposed deblurring scheme appears to be effective for the blurring problems in both conventional radiography and DTS and is applicable to improve the present image characteristics.