• Journal of KIISE:Software and Applications
• /
• v.26 no.4
• /
• pp.546-556
• /
• 1999

다층 퍼셉트론은 다양한 응용 분야에 성공적으로 적용되고 있는 대표적인 신경회로망 모델이다. 그러나 다층 퍼셉트론의 학습에 사용되는 오류역전파 알고리즘으로 알려진 기울기 강하 학습법은 느린 수렴속도로 인해 실시간 처리가 요구되거나 시간에 따라 환경이 변하는 문제에의 적용이 불가능하다. 이러한 느린 수렴속도는 기울기 강하법을 사용한 학습과정에서의 오차함수의 기울기 변화가 극히 적어 오차의 감소가 거의 일어나지 않는 부분인 플라토에 기인하는 것으로 알려져있다. 본 논문에서는 정보기하이론의 관점에서 기존의 학습법에 사용되는 기울기의 이론적 문제를 지적하고, 그로부터 플라토 문제의 원인을 밝힌다. 또한 이를 바탕으로 정보기하이론에 의해 새롭게 정의되는 자연 기울기를 이용한 학습법을 제시하고, 이를 이용한 플라토 문제가 문제해결의 가능성을 분석적으로 고찰하고 실험을 통해 확인한다.

• Transactions of the Society of Information Storage Systems
• /
• v.2 no.4
• /
• pp.236-239
• /
• 2006

Glass micro molding process is the most suitable process for fabricating high precision glass microlens amy at low cost. A new glass micro molding process was proposed. Tungsten carbide mold was fabricated by imprinting and sintering process to overcome the difficulties of the conventional process. In the glass micro molding process, process conditions such as processing temperature and compression force were changed. Geometrical properties of the replicated glass microlens array were measured and compared at variety process conditions. The condition of glass micro molding process was optimized. The experimental result showed that developed process was effective to produce a glass microlens array.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.23 no.6
• /
• pp.1425-1432
• /
• 1998

The propagation properties of Gaussian pulse signals on the microstrip line are investigated by the Kirschning & Jansen's approximated equation to meet the frequency range of a pulse, accuracy, and geometrical requiremtns of the microstrip line. The dispersion of pulse signals is analyzed regarding to the relative permittivity .epsilon.$_{r}$, substrate height h and strip width w of the microstrip line. To verify theoretical results, several samples of 50.OMEGA. microstrip lines are fabricated. And the characteristics of pulse propagation along these lines are simple measured using VNA(Vector Network Analyzer) with time-domain analysis function. A fairly good agreement has been found between the measured pulse waveforms and computer simulations.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.22 no.7
• /
• pp.1534-1542
• /
• 1997

In this paper, we present a performance improvement method for the direction-of-arrival (DOA) estimation algorithm of the narrowband signals incident on a uniform circular array. It is very important to estimate the covariance matrix effectively because the performance of DOA algorithm mainly depends on the exactness of the sampel coveriance matrix which is computed from the received samples of signals. In case of uniform circular array with the even number sensors, the structure of the arrray has a useful geometrical property. Therefore we present the method which can estimate covariance matrix more effectively using this property. The simulation results are shown to demonstrate the superior perfodrmance obtained by the proposed covariance matrix estimation method relative to that of the conventional estimation method.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.25 no.9
• /
• pp.1125-1133
• /
• 1988

This paper described construction of CIF and DXF parser using YACC (yet another compiler compiler), in order to plot CIF, and conversion of DXF into CIF for geometrical information of AUTOCAD. It is difficult to construct parper for input with complex grammar or with grammar which has possibilities of extension, and to varify the validity. YACC, a LALR parser generator, is a effective tool in these works. Also DRC or ERC can be developed quickly using primitive CIF in this paper. So, it is practicable to design IC at PC without expensive workstations.

• Transactions of the Society of Information Storage Systems
• /
• v.6 no.1
• /
• pp.25-30
• /
• 2010

In this paper, we investigate the confocal filter for high quality reading signal of micro holographic data storage system. The confocal filter is used to remove crosstalk effects of adjacent recording-layer by multi-layer multiplexing method. To remove a noise of reading signal, we consider size and geometrical structure of filter. From the obtained results, we present optical characteristics including Fresnel and Fraunhofer diffraction effects. Finally, we propose the optimal confocal filter for good image quality and increased reflection power.

• Journal of the Korea Society of Computer and Information
• /
• v.20 no.7
• /
• pp.25-32
• /
• 2015

In this paper, we propose a license plate detection method of running cars in various road images. The proposed method first classifies the road image into day and night images to improve detection accuracy, and then the tail-light regions are detected by finding red color areas in RGB color space. The candidate regions of the license plate areas are detected by using symmetrical property, size, width and variance of the tail-light regions, and to find the license plate areas of the various sizes the morphological operations with adaptive size structuring elements are applied. Finally, the plate area is verified and confirmed with the geometrical and image features of the license plate areas. The proposed method was tested with the various road images and the detection rates (precisions) of 84.2% of day images and 87.4% of night images were achieved.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2008.06a
• /
• pp.199-204
• /
• 2008

Visual surveillance is an active topic recently in Computer Vision. Event detection and recognition is one important and useful application of visual surveillance system. In this paper, we propose a new method to recognize the entering and exiting events based on the human's movement feature and the door's state. Without sensors, the proposed approach is based on novel and simple vision method as a combination of edge detection, motion history image and geometrical characteristic of the human shape. The proposed method includes several applications such as access control in visual surveillance and computer vision fields.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.25 no.9
• /
• pp.1317-1324
• /
• 2001

In this paper, a methodology of geometrical error identification and compensation for NC machine tool position is developed. We propose a reference artifact with measuring the geometry of coordinate system for compensating linear scale error of NC machine. The coordinate system of the NC machine could be compensated successfully with the information obtained by measuring the reference artifact and our compensation algorithm. Monte Carlo simulation is used to evaluate coordinate referencing ability and, the uncertainties of the machine tool position is estimated and observed through the compensation process by simulation.

• Proceedings of the IEEK Conference
• /
• 2006.06a
• /
• pp.427-428
• /
• 2006

This paper presents an image fingerprinting scheme for the print-to-capture model performed by a photo printer and digital camera. When capturing an image by a digital camera, various kinds of distortions such as noise, geometrical distortions, and lens distortions are applied. slightly and simultaneously. In this paper, we consider several steps to extract fingerprints from the distorted image in print-and capture scenario. To embed ID into an image as a fingerprint, multi-bits embedding is applied. We embed 64 bits information as a fingerprint into spatial domain of color images. In order to restore a captured image from distortions a noise reduction filter is performed and a rectilinear tiling pattern is used as a template. To make the template, a multi-bits fingerprint is embedded repeatedly like a tiling pattern. We show that the extracting is successful from the image captured by a digital camera through the experiment.