멀티코어 SoC에서 피부색상 정보와 병렬처리를 이용한 효율적인 얼굴 검출 방법

김홍희,이재흥,Kim, Hong-Hee,Lee, Jae-Heung 한국전기전자학회 2012 전기전자학회논문지 Vol.16 No.4

본 연구에서는 멀티코어에서 피부색 정보와 병렬처리를 이용하여 Viola-Jones 알고리즘을 설계하였다. 불필요한 연산을 줄이고 빠른 검출 속도를 위해 기존의 얼굴 검출 방법 중 피부색을 이용한 얼굴 검출 알고리즘을 사용하여 배경영상을 삭제하였다. 병렬처리를 위해 알고리즘을 기능별로 분할하였고, 멀티코어에서 성능을 향상시키기 위해 의존성과 크기를 고려하여 독립적으로 동작되도록 설계하였다. Cortex-A9 멀티코어가 내장된 SoC에서 실험 결과 알고리즘을 분할하기 전에 비해 약 1.8배 빠른 검출 결과를 나타내었다. In this paper, we present an implementation of Viola-Jones algorithm in a multi-core SoC by using skin color information and a parallel processing method. In order to reduce unnecessary operations and improve the detection speed, we adopted a face detection algorithm based on skin color and deleted background image. The algorithm is functionally divided into several parts taking account of the size and the dependency so that the divided functions can be proceeded in parallel. Experiment results in SoC with built-in Cortex-A9 multi core show that it is about 1.8 times faster than the existing algorithm which is not divided.

임베디드 시스템에서의 템플릿 매칭 기법을 이용한 번호판 인식 시스템 개발

김홍희,이재흥,Kim, Hong-Hee,Lee, Jae-Heung 한국전기전자학회 2011 전기전자학회논문지 Vol.15 No.4

본 연구에서는 SoC를 이용한 임베디드 시스템에 리눅스 OS 환경을 구축하고 번호판 인식 시스템을 구현하여 그 성능을 측정하였다. 자동차 번호판을 인식하기 위해서는 번호판을 검출하고 검출된 번호판을 보정 한 뒤 각 문자들에 대해 인식을 한다. 번호판 검출 방법으로는 레이블링 기법과 숫자의 특징을 이용하여 검출하였다. 검출된 번호판의 표기되어 있는 숫자들은 각각의 좌표가 있다. 이러한 숫자들의 좌표를 비교하여 영상을 보정하고 템플릿 매칭을 통해 인식을 한다. 그 결과로 번호판의 검출율은 96%, 문자 인식률은 73%, 숫자 인식률은 97%로 나타났다. 인식 시스템은 기존의 PC기반이 아닌 임베디드 보드에서 측정 되었으며 총 인식시간은 약 0.66초가 소요되었다. The implementation of the recognition system of a vehicle license plate and the Linux OS environment which is built in SoC Embedded system and its test result are presented in this paper. In order to recognize a vehicle license plate, each character has to be extracted from the whole image of a license plate and the extracted image is revised for the template matching. Labeling technique and numerical features are used to detect the vehicle license plate. Each character in the license plate has coordinates. The extracted image is revised by comparison of the numerical coordinates and recognized through template matching method. The experimental results show that the license plate detection rate is 96%, and a character recognition rate is 73%, and a number recognition rate is 97% for about 300 license plate images. The average time of the recognition in the embedded board is 0.66 sec.

프로테아제 처리가 모발의 염색성 및 형태적 특성에 미치는 영향

김홍희(Hong-Hee Kim),권태종(Tae-Jong Kwon) 한국염색가공학회 2008 韓國染色加工學會誌 Vol.20 No.2

고입력 내성을 위한 GaN HEMT 기반 S-대역 저잡음 증폭기

김홍희(Hong-Hee Kim),김상훈(Sang-Hoon Kim),최진주(Jin-Joo Choi),최길웅(Gil-Wong Choi),김형주(Hyoung-Joo Kim) 한국전자파학회 2015 한국전자파학회논문지 Vol.26 No.2

본 논문에서는 GaAs 기반 저잡음 증폭기가 사용되고 있는 레이더 수신기의 잡음지수를 낮추고, 저잡음 증폭기의 강인성(robustness)을 위하여 GaN HEMT 기반의 저잡음 증폭기를 설계하고 측정하였다. GaAs 기반의 저잡음 증폭기를 사용하는 레이더 수신기의 경우, 맨 앞단에 저잡음 증폭기를 보호하기 위한 리미터(limiter)가 필요하고, 이는 레이더 수신기 전체 잡음지수를 나빠지게 한다. 본 연구에서 측정한 GaN 기반 저잡음 증폭기의 잡음지수는 2 ㏈ 이하로 측정되었다. 상용화된 GaAs 기반 저잡음 증폭기의 경우, 최대 입력 전력은 약 30 ㏈m인 반면 본 연구에서는 입력 전력이 43 ㏈m일때 소자가 번아웃(burn-out)되었고, 전류 제한 모드로 동작시킬 경우 45.4 ㏈m에서도 강인성이 보장되었다. In this paper, we present design and measurement of LNA(Low Noise Amplifier) based on GaN HEMT(Gallium Nitride High Electron Mobility Transistor) to reduce the total noise figure of radar receiver and for robustness of LNA. In radar receiver using LNA based on GaAs(Gallium Arsenide) technology, limiter is necessary at the very front of the radar receiver to protect LNA. As a result, total noise figure of radar receiver is deteriorated. In this research, measured noise figure of LNA based on GaN HEMT is below 2 ㏈. In the case of commercialized GaAs LNA, recommended maximum input power is about 30 ㏈m. On the other hand, GaN HEMT LNA which is designed and measured is burned-out when input power is 43 ㏈m and robustness is guaranteed at input power 45.4 ㏈m.

창의 · 인성교육에 대한 초등 교사들의 인식의 개념도

김홍희(Kim, Hong-Hee),한기순(Han, Ki-Soon) 한국영재교육학회 2011 영재와 영재교육 Vol.10 No.1

The purpose of this study is to clarify elementary teachers' perceptions on creativitypersonality education by utilizing concept mapping and to examine its importance. For this, ten elementary school teachers carried out group brainstorming on creativitypersonality education, and confirmed 54 final statements through synthesizing and classifying the statements. And, based on these confirmed statements, the multi-dimensional scale and hierarchial cluster analysis using dissimilarity matrix was performed. In the analysis an average stress value of .308 was estimated. The value is appropriate for a two-dimensional point map. Also, a questionnaire survey was carried out targeting 196 elementary school teachers. The importance of elementary teachers' perceptions on creativity-personality education was analyzed. The findings are as follows: First, as a result of the hierarchical cluster analysis with the X-Y coordinate matrix, teachers' perceptions on creativity-personality were divided into four categories of ‘expected effect,’ ‘managementbased problem,’ ‘precedent condition for success,’ and ‘considerations given operation.’ Second, out of questionnaire items, the item dubbed ‘lots of results shall not be demanded in short time as for carrying out creativity-personality education.’ was indicated to have the highest importance(M=4.68). The item dubbed ‘a drop in scholastic ability is concerned with creativity-personality education.’ was indicated to have the lowest importance (M=2.87). This study aims to examine elementary teachers' perceptions on creativitypersonality education and to analyze diverse conditions about which teachers think by category in expanding and applying creativity-personality education. 본 연구의 목적은 개념도를 활용하여 창의·인성교육에 대한 초등 교사들의 인식을 밝히고, 도출된 요인들에 대하여 중요하게 지각되는 정도를 탐색해 보는 것이다. 이를 위해 창의·인성교육 교과연구회 회원으로 활동하고 있는 10명의 초등 교사가 창의·인성교육에 대해 집단 브레인스토밍을 실시하였으며, 이를 종합, 분류하는 활동을 통해 54개 최종 진술문을 확정하였다. 이들 54개의 진술문에 대한 비유사성 평정 자료를 사용하여 다차원 척도분석을 실시한 결과 2차원 개념도 제작에 적합한 stress 값은 .308이었다. 또한 196명의 초등 교사를 대상으로 설문조사하여 창의·인성교육에 대한 초등 교사들의 인식의 중요도를 분석하였다. 연구의 결과는 다음과 같다. 첫째, 개념도에 나타난 진술문의 좌표 값을 기초로 위계적 군집 분석을 실시한 결과, 창의·인성에 대한 초등 교사들의 인식의 범주는 ‘기대효과’, ‘운영상의 문제점’, ‘성공을 위한 선행조건’, ‘운영 시 고려사항’의 4개의 범주로 나타났다. 둘째, 설문문항 중 ‘창의·인성교육을 실시함에 있어 짧은 시간에 많은 실적을 요구하지 않아야 한다.’가(M=4.68) 가장 중요하게 인식되는 것으로 나타났으며, ‘창의·인성교육으로 학력의 저하가 우려된다.’가 가장 낮은 평균값(M=2.87)을 보여 상대적으로 덜 중요하게 인식되는 것으로 나타났다. 본 연구는 창의·인성교육에 대한 초등 교사들의 인식을 알아보고, 창의·인성교육을 확대·적용함에 있어 초등 교사들이 생각하는 다양한 관련 요인들을 범주별로 분석하는데 있다.

연구논문 : 식물성 염료와 검은 콩을 이용한 모발 염색에 관한 연구

김홍희 ( Hong Hee Kim ),이주영 ( Ju Young Lee ),이점숙 ( Chum Sook Lee ) 한국미용학회 2005 한국미용학회지 Vol.11 No.1

성능경의 퍼포먼스 : 다층적 경계 위의 유희

김홍희(Hong-Hee Kim) 현대미술사학회 2001 현대미술사연구 Vol.13 No.-

The artistic career of Sung NeungKyung can be divided into two phases, namely conceptual art of the 70-80s and the postmodern performance since the 90s. As his art changed from conceptual works revolving around minimalistic events and semiotic photo installations to trans-genre performance, he established a new style based on his own sensibility. However, his style can't be simply labeled as 'Sung NeungKyung style'. His work is best described as postmodern hybrid with no room for individuality, or as a pastiche that is not particulary characterizable. Seemingly being better suited for hybrid rather than pedantic conceptual art, he is continuously developing hybrid variations, which is noticeable in the sense that it clearly demonstrates his trans-genre sensibility and also it continues presently. Sung's hybrid performance is one of maximized excess; excess of actions and events, excess of props, excess of speech and text. In particular, the reading of prepared text, chanting of spells, forcing of audience participation, and other excessive chattering make his performance noisy, unfocused, tedious, gossipy, and in some instances brings to mind the mess and disorder of the market place or country theatre. Another effect of excess is boredom. His performance, which is excessive in time as well, cannot escape the bounds of boredom. It is an endless series of unrelated one-scene events and repetition of identical actions. For an audience familiar with interesting and exciting theatre, this is a new experience of boredom and puzzlement. However, boredom, a cause for aesthetic tension, can become a communicative tool that rouses the audience. In this context, the aesthetic significance of Sung's plain, tedious, and long performance that has no climax can be better understood. Sung's performance, a product of the 90s, is in fact closer to the 70s paradigm. This is because, although the pastiche and bricollage style falls under postmodernism, his work shares the utopian ideal of modern avantgarde in terms of politics. His dual identity of maintaining modern avantgarde resolve in postmodern performance is likely related to the sentiment of the 70-80s period to which he belonged. In this period dominated by monochrome painting, he had freed herself from the orbit of modernism in terms of genre as evidenced by his works with photography, installations, and events. However, in terms of aesthetics, form, and politics, these works were still conceptually linked to modernism. The fact that his performance sits on the boundary of modernism and postmodernism with a compound dual identity makes it difficult to clearly classify his art. Such difficulty in labeling her work, or the impossibility of defining traits that lacked peculiarity and individuality may have caused his art to remain in the blind spot of identity and the outskirts of art history. Or perhaps the artist, in a naive but stubborn artistic instinct, chose to remain a periphery artist with a hidden lonely joy of sitting on multiple borders.

페미니즘 비디오 미술에 나타난 나르시즘과 그로테스크

김홍희(Kim Hong-Hee) 서양미술사학회 1998 서양미술사학회논문집 Vol.10 No.-

This thesis is a study on the aspects and implications of feminist video art. It seeks to examine how feminism engages in, and operates on, the genre of video art to effect a distinctly feminist practice. The intervention of feminism in video signifies an encounter between the two major modes of postmodem discourse, video and feminism. While video may be a characteristically postmodem medium, in artistic practice, it tends to be formalizes its properties. This thesis advances the argument that it is the intervention of feminism in the contradictory condition of video art - a postmodem medium which is nevertheless amenable to the formalist logic of modernism - which enables video’s transition toward a more content-directed, socially-conscious postmodem art form. This is the primary significance of the encounter between video art and feminism, and of feminist video. The basic proposition of this thesis is the distinction between ‘aesthetic video’ and ‘political video’. Aesthetic video is related to the work that formalizes the medium’s property, while political video uses video as an information and communication tool. By the same token, feminist video can be distinguished into two categories of ‘aesthetic feminist video’ and ‘political feminist video’. The former, based on essentialist feminism, relies on aesthetics of narcissism, and the latter, derived from socialist feminism, proclaims anti/trans-narcissism. The first generation of feminist video (late 60’s - early 80’s) was grounded on the dichotomy of ‘aesthetics/politics’. On the other hand, the second generation (late 80’s ­90’s) evolved from the union of the two elements or the collapse of the dichotomy. While the first generation artists were mostly concerned with the issue of narcissism, the next generation became more focused on the sexuality and representation, but also the reinterpretation of the body extending sexual politics into body politics. It provides a powerful breakthrough for today’s feminism which attempts to deconstruct phallocentric hegemony by including ‘man trouble’ in issues of feminism. Thus, feminism video started with the conflict between aesthetics/politics, essentialism/ socialism, and narcissism/anti-narcissism, and now has proceeded into the stage of deconstructing such oppositions. This indicates the shift from narcissism to the grotesque, which is the point that this thesis attempts to clarify. In this regard, this thesis brings out narcissism and the grotesque as the main characteristics of feminist video. Narcissism is the typical aesthetics of early video art established by Rosalind Krauss and at the same time feminine aesthetics as being a frequent motif of feminist art in general. If narcissism is considered as a first generation issue, then the grotesque is referred to the second generation. The grotesque, characterized by 90’s video, performance, and body art, is identified with the feminine grotesque, and thus becomes the focal point of today’s feminist art debate. Narcissism and the grotesque are not new concepts. However, this thesis attempts to extend the two concepts by distinguishing narcissism as video aesthetics and narcissism as feminine aesthetics, and connecting the grotesque to the ‘cyborg’, respectively. In short, this thesis induces narcissism and the grotesque as the main issues of feminist video and makes an effort to extend the two concepts. At a time when women video artists are more active than ever, it seems necessary to undertake an assessment of the significance of feminist video art. It is hoped that this study will be followed by further research by others on this important subject, and also that it will lead to a new understanding of feminism and a re-examination of video art, which is now widely accepted as a genre of contemporary art. Lastly, due to practical considerations, the scope of the present research has been limited, regrettabl

DEA를 이용한 국내 신재생에너지 보급 및 기술개발 투자의 효율성 분석

김홍희(Hong-Hee Kim),이덕주(Deok-Joo Lee),김경택(Kyung-Taek Kim),박성준(Sung-Joon Park) 대한산업공학회 2014 대한산업공학회지 Vol.40 No.4

멀티코어에서 피부색상 정보와 병렬처리 방법을 이용한 얼굴 검출

김홍희 ( Hong-hee Kim ),이재흥 ( Jae-heung Lee ) 한국정보처리학회 2012 한국정보처리학회 학술대회논문집 Vol.19 No.2

최근 얼굴검출에 관한 연구는 FPGA를 통한 H/W설계부터 DSP, GPU, ARM Core에 효율적인 S/W 설계까지 다양하게 연구되고 있다. 본 연구에서는 Multi-Core에 효과적인 얼굴검출 방법을 제안한다. 피부색을 통한 얼굴 후보를 추출하고 그 외의 배경 이미지는 삭제하여 연산처리를 빠르게 하였다. Viola-Jones가 제안한 얼굴검출 알고리즘을 POSIX Thread를 사용하여 병렬 처리하였고 그 성능을 단일코어와 멀티코어에서 측정하였다. 단일 코어에서는 성능의 향상이 없었으나 멀티코어에서는 약 1.8배 속도가 향상되었고 검출 성공률은 기존과 동일하였다.