• Proceedings of The KACE
• /
• 2017.08a
• /
• pp.47-49
• /
• 2017

인공 신경망 기반 자연어 처리 시스템들에서 단어를 벡터로 변환할 때, 크게 색인 및 순람표를 이용하는 방법과 합성곱 신경망이나 회귀 신경망을 이용하는 방법이 있다. 이 때, 전자의 방법을 사용하려면 시스템이 수용 가능한 어휘집이 정의되어 있어야 하며 새로운 단어를 어휘집에 추가하기 어렵다. 반면 후자의 방법을 사용하면 단어를 구성하는 문자들을 바탕으로 벡터 표현을 생성하기 때문에 어휘집이 필요하지 않지만, 추가적인 인공 신경망 구조가 필요하기 때문에 모델의 복잡도와 파라미터의 수가 증가한다는 단점이 있다. 본 연구에서는 위 두 방법의 한계를 극복하고자 Bag of Characters를 응용하여 단어를 구성하는 문자들의 집합을 바탕으로 벡터 표현을 생성하는 방법을 제안한다. 제안된 방법은 문자를 기반으로 동작하기 때문에 어휘집을 정의할 필요가 없으며, 인공 신경망 구조가 사용되지 않기 때문에 시스템의 복잡도도 증가시키지 않는다. 또한, 단어의 벡터 표현에 단어를 구성하는 문자들의 정보가 반영되기 때문에 Out-Of-Vocabulary 단어에 대한 성능도 어휘집을 사용하는 방법보다 우수할 것으로 기대된다.

• 한국HCI학회:학술대회논문집
• /
• 2009.02a
• /
• pp.140-145
• /
• 2009

Significant attention has recently been drawn to human robot interaction system that uses face detection technology. The most conventional face detection methods have applied under pixel domain. These pixel based face detection methods require high computational power. Hence, the conventional methods do not satisfy the robot environment that requires robot to operate in a limited computing process and saving space. Also, compensating the variation of illumination is important and necessary for reliable face detection. In this paper, we propose the illumination invariant face detection that is performed under the compressed domain. The proposed method uses color balancing module to compensate illumination variation. Experiments show that the proposed face detection method can effectively increase the face detection rate under existing illumination.

• The Transactions of the Korea Information Processing Society
• /
• v.2 no.1
• /
• pp.130-136
• /
• 1995

This paper is on the development of tools to generate pointillist representation-like images automatically by computer. Pointillist representation -like effects on the generated images are enforced by steps as follows. First, the position of brush stroke is determined from the gradient vector so that the brush touches look more natural. Second, pointillist representation-like coloring is endorsed by changing saturation and value using the RGB components of image. Our approach combines image processing techniques with computer graphics techniques for more faithful pointillist representation-like images and a couple of sample images are presented to show the effectiveness.

• Journal of the Korea Computer Industry Society
• /
• v.3 no.12
• /
• pp.1765-1774
• /
• 2002

A face recognition is currently the field which many research have been processed actively. But many problems must be solved the previous problem. First, We must recognize the face of the object taking a location various lighting change and change of the camera into account. In this paper, we proposed that new method to fund feature within fast and correct computation time after scanning PC camera and ID card picture. It converted RGB color space to YUV. A face skin color extracts which equalize a histogram of Y ingredient without the luminance. After, the method use V' ingredient which transformes V ingredient of YUV and then find the face feature. The reult of the experiment shows getting correct input face image from ID Card picture and PC camera.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.37 no.4
• /
• pp.29-34
• /
• 2000

This paper presents a new indexing technique, for the fast content-based image browsing and retrieval in a database. By applying the rosette pattern that has more sample lines in the vicinity of center than those m the outer parts, we can get global gray distribution features as well as local positional information. These features are transformed into histogram and used as database indices. From the simulation results, the proposed method clearly shows the validity and the efficiency in respect of memory space as well as a good retrieval performance.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.14 no.3
• /
• pp.330-335
• /
• 2004

In this paper, we propose the technique for recognizing the human emotion by using the color image. To do so, we first extract the skin color region from the color image by using HSI model. Second, we extract the face region from the color image by using Eigenface technique. Third, we find the man's feature points(eyebrows, eye, nose, mouse) from the face image and make the fuzzy model for recognizing the human emotions (surprise, anger, happiness, sadness) from the structural correlation of man's feature points. And then, we infer the human emotion from the fuzzy model. Finally, we have proven the effectiveness of the proposed method through the experimentation.

• Journal of Sensor Science and Technology
• /
• v.26 no.6
• /
• pp.420-426
• /
• 2017

A colorimetric sensor was investigated to achieve a low-cost warning device for harmful gaseous formaldehyde (HCHO). The sensor is based on selective reactions between hydroxylamine sulfate and HCHO, leading to the production of sulfuric acid. The produced acid results in color-changing response through the acid-base reaction with dye molecules impregnated on a solid membrane substrate. For attaining this purpose, sensors were fabricated by drop-casting a dye solution prepared using different pH indicators on various commercially available polymer sheets, and their colorimetric responses were evaluated in terms of sensitivity and reliability. The colorimetric sensor using bromophenol blue (BPB) and nylon sheet was found to exhibit the best performance in HCHO detection. An initial bluish green of a sensor was changed to yellow when exposed to gaseous formaldehyde. The color change was recorded using an office scanner and further analyzed in term of RGB distance for quantifying sensor's response at different HCHO(g) concentrations. It exhibited a recognizable colorimetric response even at 50 ppb, being lower than WHO's standard of 80 ppb. In addition, the sensor was found to have quite good selectivity in HCHO detection under the presence of common volatile organic compounds such as ethanol, toluene, and hexane.

• The Journal of Information Technology
• /
• v.5 no.1
• /
• pp.99-111
• /
• 2002

Due to non-linear characteristics of color spaces, color corrections using linear conversions for real image near color reappearance causes color distortions. In order to overcome this problem, the Bezier Curve, constructed with a set of arbitrary plane in the linear theory, has been used. However, the Bezier Curve increases in proportion to the number of data points, resulting in higher computational complexities. This paper attempts to use a Triplicated Piecewise Bezier Cubic-Curve (TPBC-Curve) of which the degree is cubic on the whole interval while keeping the characteristics of Bezier Curves. By Comparing the TPBC-Curve with Bezier Curve of 20 degree, the paper not only reduces the distortion during color correction but also lessens the relative increase of workload that is caused by the color correction in a small zone.

• Journal of Korea Game Society
• /
• v.8 no.3
• /
• pp.69-76
• /
• 2008

This paper presents a novel rendering algorithm based on sperical coordinate representation of the object. The vertices of the object are transformed into the sperical coordinate system, and we construct additional maps: the centroid and index of the triangle, the memory access table. While OpenGL rendering pipeline touches all vertices of an object, the proposed method takes account of the only visible vertices by examining the visible triangles of the object. Simulation results demonstrated that the proposed method achieve an efficient rendering performace.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2014.11a
• /
• pp.77-80
• /
• 2014

멀티미디어 기술의 보급 확산 및 급속한 발전으로 새로운 영상 압축 기술인 HEVC(High Efficiency Video Coding) 고화질 영상 압축 표준을 탄생시켰으며, 사용자 또한 대형 화면에 대한 선호도가 높아지고 있다. 그 결과 기존의 HD급 영상보다 4배 이상, 16배까지 선명한 초고화질 UHD(Ultra High Definition) 영상 서비스가 차세대 방송기술로 새롭게 주목받고 있다. 또한 JPEG 2000 압축도 기존 처리된 $4096{\times}4096$ 픽셀 이미지를 넘어 초고화질 해상도 이미지(8K : $7680{\times}4320$ 혹은 $8192{\times}4320$ 픽셀)를 처리 지원을 하고 있다. 따라서 초고화질 이미지의 획득 및 저장을 위해서는 고속의 처리 기술이 필요하다. 이에 본 논문은 초고화질 해상도 이미지의 고속 처리를 위한 병렬처리 기술에 대해 단계적 연구를 실행하며, 이를 위하여 1차적으로. JPEG 2000의 처리 과정을 살펴보고 전처리 단계인 색공간 변환 알고리즘 적용을 위하여 사용자 정의의 쓰레드 기반 고속처리를 수행하였다. 실험 결과 기존의 처리보다 사용자 정의 기반 쓰레드 고속처리가 초고화질 해상도 이미지(UHD 8K : $7680{\times}4320$)를 기준으로 최대 15배의 성능 향상의 결과를 보여주었다.