• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2005.11a
• /
• pp.191-194
• /
• 2005

본 논문에서는 방송 영상에서 조명효과와 크기변화 등에 강인한 얼굴패턴 검출기법을 제시한다. 제안된 얼굴검출 모델은 영상 전처리 과정과 얼굴패턴 검출 과정으로 이루어진다. 전처리 과정은 조명변화에 대한 보정기능과 다중필터에 의한 후보영역 선별기능으로 구분된다. 얼굴패턴 검출과정은 다단계의 특징지도 생성과정과 패턴분류 과정으로 이루어진다. 특징지도를 생성하기 위하여 가보(Gabor) 필터계층을 포함하는 CNN(Convolutional Neural Networks)모델을 도입하였다. 다양한 배경을 고려한 효과적인 학습을 위하여 본 논문에서는 억제성의 뉴런(Inhibitory neuron)을 포함하는 구조의 CNN모델을 적용한다. CNN으로부터 추출되는 특징집합은 최종 단계에서 WFMM(Weighted Fuzzy Min Max) 모델을 사용하여 분류된다. 이때 사용되는 특징집합의 크기는 분류기의 규모 및 계산량의 결정적인 역할을 준다. 이에 본 연구에서는 최종 분류 과정에 사용되는 특징의 수를 효과적으로 줄이기 위해 FMM모델을 사용하는 적응적인 특징 선별 기법을 제안한다. 또한 실제 영상을 통한 실험결과로부터 제안된 이론의 타당성을 고찰한다.

• Journal of the Korea Society of Computer and Information
• /
• v.19 no.6
• /
• pp.11-17
• /
• 2014

While safe and convenient, ultrasound imaging analysis is often criticized by its subjective decision making nature by field experts in analyzing musculoskeletal system. In this paper, we propose a new automatic method to extract muscle area using ART2 neural network based quantization. A series of image processing algorithms such as histogram smoothing and End-in search stretching are applied in pre-processing phase to remove noises effectively. Muscle areas are extracted by considering various morphological features and corresponding analysis. In experiment, our ART2 based Quantization is verified as more effective than other general quantization methods.

• Journal of Korea Water Resources Association
• /
• v.35 no.2
• /
• pp.195-201
• /
• 2002

In this study, an equation is proposed to estimate the limit velocity for lateral stable bed in a curved channel stream. The stable bed on lateral direction is satisfied when there is no more deformation occurs on the transverse bed slope and non-scouring condition in a bend. A theoretical equation for limit velocity is derived using a transverse bed slope model. So, the limit velocity has its theoretical background in the equilibrium of two forces, lateral shear force at the bed due to longitudinal flow and the corresponding lateral bed shear force. To verify the equation, data from four natural river channels were used. There is good agreement between the calculated values using this equation and the measured values. The corrections in equation was found to be correlated with the averaged particle Froude number.

• Proceedings of the Korea Contents Association Conference
• /
• 2008.05a
• /
• pp.15-18
• /
• 2008

This paper proposes a method to reconstruct the 3-D face using structured light image. First of all, we suppose that each sight vector of a projector and camera are parallel. We project the structured light in the shape of lattice on the background to acquire the reference-structured light image. This image is used to calibrate the projector and camera. Since then, we acquire the face-structured light image which is projected the same structured light on the face. These two structured light images are used to reconstruct the 3-D face through the variation which is measured from the positional difference of feature vectors. In our experiment result, we could reconstruct the 3-D face image as recognize through these simple devices.

• Journal of Korea Multimedia Society
• /
• v.12 no.8
• /
• pp.1082-1088
• /
• 2009

Fundus images are highly useful in evaluating patients' retinal conditions in diagnosing eye diseases. In particular, vessel regions are essential in diagnosing diabetes and hypertension. In this paper, we used top-hat filter to compensate for non-uniform background. Image contrast was enhanced by using contrast limited adaptive histogram equalization (CLAHE) method. Hessian matrix was next applied to detect vessel regions. Results indicate that our method is 1.3% more accurate than matched filter method. Our proposed method is expected to contribute to diagnosing eye diseases.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2019.05a
• /
• pp.63-63
• /
• 2019

최근 기후변화로 인한 도시지역 침수 피해를 저감하기 위한 다양한 연구가 수행되고 있으며 침수해석을 위한 기초자료로써 GIS 기반 하수관망 자료 활용의 중요성이 증대되고 있다. 그러나 이러한 하수관망 자료의 대부분은 지자체 수준의 행정단위에 의해 작성/관리 되고 있으며 하수관 망의 유지보수에 중점을 두어 제작되었기 때문에 침수해석을 위한 속성자료가 누락되어 있는 경우가 상당수 존재한다. 따라서 고유의 제작 목적과 침수해석이라는 활용 목적이 일치 하지 않아 속성 데이터 값이 존재하지 않거나 침수 모델링에 필요한 필수 정보가 누락되어 개별 연구자들이 별도의 보완작업을 수행한 후 침수해석에 활용하고 있는 실정이다. 이러한 개인연구자들의 주관적 판단에 의한 하수관망의 단순화 또는 보완작업은 상황에 따라 자료의 불확실성을 증대시키며 연구자의 숙련도와 배경지식에 따라 침수 해석 결과에 많은 영향을 미치고 있다. 따라서 GIS기반 하수관망 자료를 침수 모의에 활용 가능한 입력 자료로 변환 하는 경우 개별 연구자들의 주관적 개입이 최대한 배제된 형태의 자료를 만들기 위한 기본 알고리즘 개발이 시급한 상태이다. 본 연구에서는 서울시 사당역 인근 유역과 부산시 온천천 유역의 GIS 기반 하수관망 자료의 형식에 대해서 알아보고 누락 자료를 보완하기 위한 알고리즘을 개발하였다. 개발된 알고리즘을 활용하여 누락자료가 보완된 하수관망 자료는 향후 개별 연구자들의 주관적 판단을 배제하여 도시침수 해석 시 하수관망 자료의 불확실성을 최소화 하는데 기여할 수 있을 것으로 판단된다.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2009.05a
• /
• pp.364-368
• /
• 2009

자궁경부암은 다른 암과 달리 전암(前癌) 단계가 존재하므로 조기에 발견할 경우 생존율이 높다. 그러나 검체 적정성의 부족과 검체 체취의 오류로 인해 질병이 있음에도 음성으로 나타나는 위음성률이 높다. 따라서 본 논문에서는 세포 도말검사에서 사용되는 자궁 경부진 세포에서 암종 세포를 추출하는 방법을 제안한다. 영상의 배경 그리고 핵과 세포질 영역의 구분이 중요하기 때문에 조기 자궁 경부 세포진 영상에서 핵의 추출은 Lighting Compensation을 적용하여 영상을 보정하고, 명암도 분포가 가장 작은 B 채널과 명암도 분포가 높은 R채널과의 OR 연산을 적용한 후, $3{\times}3$마스크를 이용하여 잡음을 제거한다. 잡음이 제거된 영상을 이진화하고 Grassfire 알고리즘을 이용하여 암종 세포의 후보 객체를 추출한다. 추출된 세포 객체에서 핵의 크기, 핵의 면적과 핵의 외곽의 방향성 정보를 이용하여 백혈구와 잡음으로 구성된 객체를 제거한다. 세포 도말검사 과정에서 겹쳐진 부분은 거리 함수와 명암도를 이용하여 마커를 추출하고 추출된 마커 정보와 워터쉐드 알고리즘을 적용하여 겹쳐진 암종 세포를 분리한다. 자궁경부 편평 세포진 400 배율 영상과 자궁 경부 상피내 종양 400 배율 영상을 대상으로 실험한 결과, 기존의 자궁 경부진 암종 세포 추출 방법보다 효과적으로 암종 세포 영역이 추출되는 것을 확인하였다.

• Geophysics and Geophysical Exploration
• /
• v.7 no.3
• /
• pp.164-173
• /
• 2004

To delineate geothermal water movement at the Pohang geothermal development site, Self-Potential (SP) survey and monitoring were carried out during pumping tests. Before drilling, background SP data have been gathered to figure out overall potential distribution of the site. The pumping test was performed in two separate periods: 24 hours in December 2003 and 72 hours in March 2004. SP monitoring started several days before the pumping tests with a 128-channel automatic recording system. The background SP survey showed a clear positive anomaly at the northern part of the boreholes, which may be interpreted as an up-flow Bone of the deep geothermal water due to electrokinetic potential generated by hydrothermal circulation. The first and second SP monitoring during the pumping tests performed to figure out the fluid flow in the geothermal reservoir but it was not easy to see clear variations of SP due to pumping and pumping stop. Since the area is covered by some 360 m-thick tertiary sediments with very low electrical resistivity (less than 10 ohm-m), the electrokinetic potential due to deep groundwater flow resulted in being seriously attenuated on the surface. However, when we compared the variation of SP with that of groundwater level and temperature of pumping water, we could identify some areas responsible to the pumping. Dominant SP changes are observed in the south-west part of the boreholes during both the preliminary and long-term pumping periods, where 3-D magnetotelluric survey showed low-resistivity anomaly at the depth of $600m\~1,000m$. Overall analysis suggests that there exist hydraulic connection through the southwestern part to the pumping well.

• Journal of KIISE:Computing Practices and Letters
• /
• v.8 no.4
• /
• pp.479-488
• /
• 2002

In recent years the amount of digital video used has risen dramatically to keep pace with the increasing use of the Internet and consequently an automated method is needed for indexing digital video databases. Textual information, both superimposed and embedded scene texts, appearing in a digital video can be a crucial clue for helping the video indexing. In this paper, a new method is presented to extract both superimposed and embedded scene texts in a freeze-frame of news video. The algorithm is summarized in the following three steps. For the first step, a color image is converted into a gray-level image and applies contrast stretching to enhance the contrast of the input image. Then, a modified local adaptive thresholding is applied to the contrast-stretched image. The second step is divided into three processes: eliminating text-like components by applying erosion, dilation, and (OpenClose+CloseOpen)/2 morphological operations, maintaining text components using (OpenClose+CloseOpen)/2 operation with a new Geo-correction method, and subtracting two result images for eliminating false-positive components further. In the third filtering step, the characteristics of each component such as the ratio of the number of pixels in each candidate component to the number of its boundary pixels and the ratio of the minor to the major axis of each bounding box are used. Acceptable results have been obtained using the proposed method on 300 news images with a recognition rate of 93.6%. Also, my method indicates a good performance on all the various kinds of images by adjusting the size of the structuring element.

• Journal of Korea Water Resources Association
• /
• v.45 no.2
• /
• pp.203-215
• /
• 2012

The objective of this study is to develop a catchment hydrologic cycle assessment model which can assess the impact of urban development and designing water cycle improvement facilities. Developed model might contribute to minimize the damage caused by urban development and to establish sustainableurban environments. The existing conceptual lumped models have a potential limitation in their capacity to simulate the hydrologic impacts of land use changes and assess diverse urban design. The distributed physics-based models under active study are data demanding; and much time is required to gather and check input data; and the cost of setting up a simulation and computational demand are required. The Catchment Hydrologic Cycle Assessment Tool (hereinafter the CAT) is a water cycle analysis model based on physical parameters and it has a link-node model structure. The CAT model can assess the characteristics of the short/long-term changes in water cycles before and after urbanization in the catchment. It supports the effective design of water cycle improvement facilities by supplementing the strengths and weaknesses of existing conceptual parameter-based lumped hydrologic models and physical parameter-based distributed hydrologic models. the model was applied to Seolma-cheon catchment, also calibrated and validated using 6 years (2002~2007) hourly streamflow data in Jeonjeokbigyo station, and the Nash-Sutcliffe model efficiencies were 0.75 (2002~2004) and 0.89 (2005~2007).