• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.32 no.4
• /
• pp.241-248
• /
• 2008

The purpose of the research is to investigate of diesel spray combustion for simultaneously reduction way of NOx and PM. The diesel injection were done into intermediates that are generated by very lean DME HCCI combustion using a RCM. The concentration of intermediate could not be directly measured, so we estimated it by CHEMKIN calculation. Two dimensional spontaneous luminescence images which are created by chemical species reaction at low temperature reaction (LTR) and high temperature reaction (HTR) are captured by using a framing streak camera. Also, combustion events were observed by high-speed direct photography. The ignition and combustion events were analyzed by pressure profiles and the KL values and flame temperatures were analyzed by the two-color method.

• Journal of the Korea Computer Industry Society
• /
• v.9 no.3
• /
• pp.129-136
• /
• 2008

This paper implement image gradient detection algorithm with high-performance DSP. First the NTSC color image convert to B/W image. The image gradient detect with Hough transform after edge detection image from the B/W images. The value of image gradient detection control the servo motor to original position of the NTSC camera if camera base to the left or right tilt.

• Proceedings of the Korea Multimedia Society Conference
• /
• 2000.11a
• /
• pp.266-270
• /
• 2000

내용기반 영상검색시스템에서 객체 단위로 영상을 검색하기 위해서는 영상에서 의미있는 객체를 추출하는 과정이 필수적이며, 이를 위해 영역 분할을 효율적으로 수행하기 위한 양자화가 선행되어야 한다. 일반적인 칼라 양자화 기법은 칼라 수를 줄이되 양자화 된 영상이 원시 영상과 가능할 비슷해 보이도록 하는 것을 목적으로 하지만, 영역 분할을 위한 칼라 양자화에서는 칼라의 표현보나는 의미있는 객체를 용이하게 추출할 수 있도록 양자화 하는 것을 목적으로 한다. 본 논문에서는 기존의 Octree 양자화 방법과 K-means 알고리즘의 장점을 조합하여 영역 분할에 용이한 양자화 결과를 얻을 수 있는 방법을 제안한다. 먼저, Octree 양자화 방법을 수행하여 얻어진 양자화 된 칼라들 중에서 시각적으로 유사한 칼라를 병합함으로써, Octree 양자화 방법의 단점인 강제 분할 문제점을 해결한다. 이어서, 병합 후의 양자화 된 칼라에 대해서만 K-means 알고리즘을 수행함으로써, 보다 빠른 시간 내에 영역 분할에 적합한 양자화 된 영상을 얻는다. 실험을 통해 제안한 방법의 효용성을 확인하였다.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2008.10a
• /
• pp.503-506
• /
• 2008

The integral imaging is a promising 3D display technology since it is able to deliver continuous viewing points, full parallax, and full color view to the observers in space. In this paper, we propose a novel interactive 3D integral imaging system using a single camera. The user interface is implemented by adding a camera in the conventional integral imaging system. To show the possibility of the proposed system, we implement the optical setup and present the preliminary results

• Journal of the Korean Society of Costume
• /
• v.54 no.2
• /
• pp.39-52
• /
• 2004

The purpose of this comparative study what is the way to express difference fashion design between Japanese fashion designers and Western fashion designers in the 1990s. The design data were 422 pictures of contemporary fashion designs which represented the Japanism style design from fashion collections and analyzed by literature study and classification categories. The results were as follows : 1) Early general Japanism expression was directly borrowed from traditional fashion design elements and has changed indirectly with fashion trend. 2) Japanese fashion designers have used color, fabric, and pattern than shape to express Japanese traditional images. 3) Western fashion designers have directly borrowed aesthetic elements of Kimono design, that is, layering. bending, neck-line, sleeve, and traditional Japanese men's wear to express Japanism.

• International Journal of Costume and Fashion
• /
• v.7 no.2
• /
• pp.32-40
• /
• 2007

American trendy casual styles can be characterized as fitted to body and rendering sexy and fashionable images and now very popular in Korea. This paper aims to identify diffusion channels of Amreican trendy casual style among Korean market and the characteristics of imitation for American trendy casual styles made by Korean casual brands. American trendy casual style was introduced and diffused mainly through 3 channels: American films and TV dramas, on-line communities, on-line shopping mall speciallized for purchasing American apparel brands. Some of American celebrities became fashion ikons and played critical roles in diffusing American trendy styles. After searching websites and fashion magazines and visiting brand stores, several Korean casual brands imitated the styles of American trendy casual brands and the scope of imitation was vast across items. Imitation was found mainly among T-shirts,jacket, jumper, pants, skirt and accessaries. Also imitation was found in almost every aspects of clothes such as shape, sihlouette, color, prints and stitching, belt, and drawstring.

• Proceedings of the KSME Conference
• /
• 2000.11a
• /
• pp.559-563
• /
• 2000

In this paper we describe a method for aligning a robot gripper using image information. The region of gripper is represented from HSI color model that has major advantage of brightness independence. In order to extract the feature points for vision based position control, we find the corners of gripper shape using polygonal approximation method which determines the segment size and curvature of each points. We apply the vision based scheme to the task of alignment of gripper to reach the desired position by 2 RGB cameras. Experiments are carried out to exhibit the effectiveness of vision based control using feature points from polygonal approximation of gripper.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.26 no.8B
• /
• pp.1161-1167
• /
• 2001

본 논문에서는 칼라 영상에 대한 인간 시각 특성과 무채색으로부터 정보를 추출하여 칼라 정지 영상에 효과적으로 워터마크 신호를 삽입하는 방법을 제안하였다. 워터마크는 이진 워터마크를 사용하였으며, 일반적으로 인간 시각 체계 특성에서 무채색(achromatic) 신호보다 유채색(chromatic) 신호에 대한 변별력이 떨어지는 특성을 이용하여 입·출력 장치로 획득된 RGB 좌표계의 칼라 정지 영상을 무채색 성분과 유채색 성분으로 완벽히 분리되는 다른 좌표계로 변환하여 워터마크를 유채색 포화도 성분에 삽입한다. 워터마크를 검출하는 과정은 원영상을 사용하지 않고 유채색 포화도 성분의 변경 형태를 추정하여 삽입된 워터마크를 검출 가능하였다. 실험 결과 특히 비가시성 특성 부분에서는 기존 방법들과 비교할 때 향상된 결과를 얻을 수 있었으며 강인성 또한 좋은 결과를 얻었다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2007.10a
• /
• pp.229-230
• /
• 2007

Lane detection is one of the major components of traffic intelligence. It is impossible to recognize lanes as human do in all kinds of special situations; however, we can try to solve special problems with special methods. In this paper we propose a simple method using color segmentation, the Probabilistic Hough Transform (PHT), and the Least-Square in real-time lane tracking. Vehicles in neighborhood can be eliminated with one simple threshold in segmentation. Meanwhile, broken shape lanes in different road conditions can be successfully detected using the combination of PHT and Least-Square method. Eventually, this method is tested with groups of static images downloaded from internet and video sequences shot randomly on some highways. Satisfactory results are received.

• Proceedings of the Korean Information Science Society Conference
• /
• 1998.10c
• /
• pp.455-457
• /
• 1998

동영상에서의 얼굴 영역 추출은 헤드 제스처 인터페이스를 위한 기본적이고 필수적인 기법이다. 얼굴 영역 추출을 위해서는 색상 정보와 차영상을 이용한 방법이 많이 사용되며, 색상 정보를 이용하는 방법에는 HSI의 H(hue)성분과 YIQ의 I(in-phase)성분이 널리 알려져 있다. 본 논문에서는 먼저 얼굴 색상에 해당하는 각 색상 성분의 구간을 탐색하고, 다음으로 각 색상 정보를 이용한 얼굴 영역 검출의 정확도를 비교 실험한다. 또한, 색상 정보와 차영상을 결합한 방법에 대해서도 얼굴 영역 검출의 정확도를 비교한다. 실험 결과, YIQ의 경우 구간 130~150, HSI의 경우 구간 0~20에서 얼굴색을 잘 표현하는 것으로 나타났다. 얼굴 영역 검출의 정확도 측면에서는, 색상 정보만을 이용한 실험의 경우 YIQ가 HSI에 비해 약 10%의 향상된 성능을 보였고, 색상 정보와 차영상을 결합한 경우에서도 YIQ가 약 5%의 향상된 성능을 보였다.