• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.47 no.1
• /
• pp.152-159
• /
• 2010

Existing character recognition algorithm recognize characters in simple conditional. It has the disadvantage that recognition rates often drop drastically when input document image has low quality, rotated text, various font or size text because of external noise or data loss. In this paper, proposes the optical character recognition algorithm which using bicubic interpolation method for the catalog retrieval when the input image has rotated text, blurred, various font and size. In this paper, applied optical character recognition algorithm consist of detection and recognition part. Detection part applied roberts and hausdorff distance algorithm for correct detection the catalog of book. Recognition part applied bicubic interpolation to interpolate data loss due to low quality, various font and size text. By the next time, applied rotation for the bicubic interpolation result image to slant proofreading. Experimental results show that proposal method can effectively improve recognition rate 6% and search-time 1.077s process result.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.46 no.2
• /
• pp.92-99
• /
• 2009

This paper established the range of the wavelength of traffic lights to detection the color of traffic lights and the color component segmentation with the range of the wavelength. Development of traffic light automatic discrimination system is consists of the color detection and the traffic lights recognition. In this thesis, it established the range of the wavelength of traffic lights to detection the color of traffic lights and the color segmentation with the range of the wavelength. By the segmentation, the traffic light colors(red, orange and green) can be detected and the background is changed into gray image. Next, we proposed the algorithm which can detect the area of traffic lights in the various surroundings with the wavelet transformation algorithm. Also, we proposed traffic lights recognition algorithm using between the edge operator and the Hausdorff distance algorithm based on CBIR(Content-based Image retrieval). Therefore, the proposed algorithm is more superior to the conventional algorithm by experimenting with the illumination including the traffic lights and the backgrounds with various images.

• Journal of the Korea Computer Graphics Society
• /
• v.25 no.3
• /
• pp.93-103
• /
• 2019

We introduce a geometric processing method for freeform surfaces based on high-precision torus patch approximation, a new spatial data structure for efficient geometric operations on freeform surfaces. A torus patch fits the freeform surface with flexibility: it can handle not only positive and negative curvature but also a zero curvature. It is possible to precisely approximate the surface regardless of the convexity/concavity of the surface. Unlike the traditional method, a torus patch easily bounds the surface normal, and the offset of the torus becomes a torus again, thus helps the acceleration of various geometric operations. We have shown that the torus patch's approximation accuracy of the freeform surface is high by measuring the upper bound of the two-sided Hausdorff distance between the freeform surface and set of torus patches. Using the method, it can be easily processed to detect an intersection curve between two freeform surfaces and find the offset surface of the freeform surface.

• Journal of the Korean Society of Industry Convergence
• /
• v.14 no.3
• /
• pp.105-111
• /
• 2011

Landmark Spots in 2D gel electrophoresis are used in many methods of 2DEG spot matching. Landmark Spots are obtained manually and it is a bottle neck in the entire protein analysis process. Automated landmark spots detection is a very crucial topic in processing a massive amount of 2DGE data. In this paper, Automated landmark spot detection is proposed using point pattern matching and graph theory. Neighbor spots are defined by a graph theory to use and only a centered spot and its neighbor spots are considered for spot matching. Normalized Hausdorff distance is introduced as a criterion for measuring degree of similarity. In the conclusion, the method proposed in this paper can get about 50% of the total spot pairs and the accuracy rate is almost 100%, which the requirements of landmark spots are fully satisfied.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.18 no.2
• /
• pp.193-199
• /
• 2008

In this paper, it is proposed the implementation of face authentication system based on real-time image processing. We described the process implementing the two steps for real-time face authentication system. At first face detection steps, we describe the face detection by using feature of wavelet transform, LoG operator and hausdorff distance matching. In the second step we describe the new dual-line principal component analysis(PCA) for real-time face recognition. It is combines horizontal line to vertical line so as to accept local changes of PCA. The proposed system is affected a little by the video size and resolution. And then simulation results confirm the effectiveness of out system and demonstrate its superiority to other conventional algorithm. Finally, the possibility of performance evaluation and real-time processing was confirmed through the implementation of face authentication system.

• Proceedings of the KIEE Conference
• /
• 2005.10b
• /
• pp.539-541
• /
• 2005

We propose efficient video retrieval scheme which use fast key frame extraction in DCT domain. Our scheme extract key frame using the edge histogram difference which is extracted in compressed domain for I-frames. And the video retrieval is implemented using Hausdorff distance function about edge histogram of key frame. This approach enables fast content-based video retrieval of the compressed video content without decompression process. Experimental results show our scheme is very fast and efficient.

• Proceedings of the IEEK Conference
• /
• 2003.07e
• /
• pp.2104-2107
• /
• 2003

In this paper, we propose the method Model based Non-Rigid Moving Object Tracking. Motion based method becomes difficult to predict precisely when motion gets larger, so that we can solve such difficultly with regarding the moving object as a model. In the model based method, it should be concerned about setting initial model and updating its model in each frame. We used SNAKE in a way to set the initial model, and also proposed a modified SNAKE to handle the previous SNAKE problems. Moreover, with the elliptical setting, we made the initializing process automatically which is highly subject to change in measuring the performance of SNAKE. We used the Hausdorff distance to identify models in each frame. Through our experiments, our Proposed algorithm does effective work in Non-Rigid Moving Object Tracking.

• Journal of the Korean Mathematical Society
• /
• v.34 no.3
• /
• pp.695-706
• /
• 1997

We investigate the bahivor of the gradient of solutions to the refraction equation $div(( 1+ (k - 1)_\chi D)\nabla u) = 0$ under perturbation of domain D. If $u_h$ are solutions to the refraction equation corresponding to subdomains D and $D_h$ of a domain $\Omega$ in 2 dimensional plane with the same Neumann data on $\partial\Omega$, respectively, we prove that $\left\$\mid$ \nabla(u - u_h) \right\$\mid$_{L^2(\Omega)} \leq C\sqrt{dist(D, D_h)}$ where $dist(D, D_h)$ is the Hausdorff distance between D and $D_h$. We also show that this is the best possible result.

• Proceedings of the Korean Information Science Society Conference
• /
• 2001.04b
• /
• pp.526-528
• /
• 2001

기존의 얼굴표정 인식연구의 대부분은 얼굴영상에서 사전정보 획득과, 인식이 각각 별개로 수행되어, 전자의 결과가 후자를 보장하지 못하거나, 데이터와 계산 양의 과다, 그리고 인지과정이 사람과 다르다는 등의 문제가 있다. 이에 대해 하우스도르프 거리 매칭을 적용, 표정인식을 시도한다. 이는 전체적인 유사도를 측정하는 방법으로서 전체이론(Holistic theory)에 기반하여, '사람의 인지과정'을 따른다. 그러나 축소된 데이터를 사용하므로, 이 방법의 인식결과가 부족할 경우, 영상워핑을 적용하여 Brennan과 Carton이 제안한 캐리커쳐 효과를 이용한다. 이는 영상을 적절히 변형, 표정의 특징을 과장하고 잡영을 제거하여, 인식하기 쉬운, 분명한 표정을 생성하는 방법이다. 위 과정을 통해, 사람의 인지과정을 모사하고, 최소한의 데이터로써 사전정보 획득과정이 생략된, 입력영상으로부터 직접 표정을 인식하는 방법을 제안한다.

• Journal of the Korean Society for Industrial and Applied Mathematics
• /
• v.17 no.3
• /
• pp.171-179
• /
• 2013

In this paper we present a method of circular arc approximation by quartic B$\acute{e}$zier curve. Our quartic approximation method has a smaller error than previous quartic approximation methods due to the alternation of the error function of our quartic approximation. Our method yields a closed form of error so that subdivision algorithm is available, and curvature-continuous quartic spline under the subdivision of circular arc with equal-length until error is less than tolerance. We illustrate our method by some numerical examples.