
http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
李泰鴻 嶺南大學校 工業技術硏究所 1986 연구보고 Vol.14 No.1
The poles appear to be the peaks in the frequency spectrum characteristics diagram. For the poles that are placed nearby with each other, the peakes become mingled together to make one peak as a whole. Sometimes we need to know the frequency of a pole to set up the system configuration. By adjusting the radius of circle that counts for the frequency response of the system, we can separate the mingled peak frequency on the z-plane.
李泰鴻 嶺南大學校 工業技術硏究所 1988 연구보고 Vol.16 No.1
Before we process the procedure to recognize the printed characters, it is necessary to preprocess their images obtained by the optical systems. Non-ideal point spread function of the optical systems blurs their images. Deblurring the character images in the spatial domain can be processed using the van Cittert recursive method. Psf, in general, involves the negative portions whose repeated applying reverses the phase each time. The phase reversal causes oscillations in th deblurred images. This can be corrected by deblurring 2nd times for each recursion, which reverses the phase 180°once more to become 0° phase. This phase corrected van Cittert algorithm in the spatial domain makes it possible to simplify the method and to reduce the processing time compared with that of the frequency domain process.
조영창,이태홍 嶺南大學校 工業技術硏究所 1997 工業技術硏究所論文集 Vol.25 No.2
In image restoration for the severe degradation of noise, the solution spaces which have many local optima around global solution appear. Thus the algorithms which find the global optima among many local optima in the solution space are superior to the general restoration technique like gradient methods. We introduce genetic algorithm is this paper and restore the image degraded by severe noise. And we proposed genetic operators, crossover and mutation, suited for gray scaled images. Finally we also proposed the local fine tuning of window size and visit number for delicate searching mechanism in the vicinity of the global solution. Through the experimental results we verified the effectiveness of the proposed genetic operators on the gray scaled images over the conventional ones, as well as the improved performance of local fine tuning.
이태홍 嶺南大學校 工業技術硏究所 1992 연구보고 Vol.20 No.1
An iterative image restoration is introduced that processes with varying iteration numbers according to the variance of the neighborhood pixels. In general the other method is applied to the whole image without considering the local pixel informations, which is not so effective for the processing time. So in this paper we process local iterative restoration method which iterates many times on detailed portion while in the flat region the iteration decreases. Local iteration method can use any iteration method because it varies in the iteration numbers only. not convergence, which shows decreased processing time.
선형 Filter와 비선형 Filter 조합에 의한 화상복원
李泰鴻 嶺南大學校 工業技術硏究所 1986 연구보고 Vol.14 No.1
In the image restoration, the smoothing filter to reduce the effect of white Gaussian noise can be used. For the image of lower noise it acts well. But for the higher noise iterative smoothing is needed. Each time an iteration is processed, some area shows better result, while the others worse. Linear filter usually blurrs the edge portion, but it shows better effect for the flat plateau or back ground. Nonlinear filter shows excellent effect for preserving edges. Severe noise makes the both filters helpless. Iteration is needed in that case. The iterated linear filter makes the edge blurred excessively, while the iterated nonlinear filter makes streaks in line because of the propagated error. Combinations of both types of filter will have superior performance to that of filter used separately.
이태홍 한국통신학회 1992 한국통신학회논문지 Vol.17 No.8
A local iterative Image restoration method Is Introduced that processes with varying iteration numbers according to the local statistics. In general almost of the Iterative method applies Its algorithm to the whole Image without considering the local pixel informations, which Is not so effective for the processing time. Usually the edges or details have an Important role In visual effect. So in this paper we process the edges or the details many times while In the flat region we just pass over or decrease iterations. This method shows good MSE (Mean Square Error) improvement, better subjective qualify and reduced processing time.
휘발성 데이타 메모리의 불휘발성 프로그램 메모리化에 관한 一方案
李泰鴻 嶺南大學校 工業技術硏究所 1982 연구보고 Vol.10 No.2
A simple and general purposed nonvolatile read-write-memory system applicable to any circuit in place of EAROM, which is very complicated and difficult to use because of its diverse voltages, timing and control signals, is suggested in this paper. Using CMOS RAM and rechargeable battery back up, we can modify easily the data table or even the program memory in circuit without extracting memory devices. It is convenient for the equipments containing time variant parameters or data in its program. It is also examined how long the data can be retained when the back up supply recharged for once. Since this system recharges the back up supply whenever the main voltage source is used, the retention of data is almost everlasting.