신경회로망과 k-NN법의 식별능력비교 : 모음 및 膝蓋骨亞脫臼CT화상인식에 있어서의 고찰 A Study in Vowel and Patellar Subluxation Image Recognitions

金應珪,權寧島 大田産業大學校 1994 한밭대학교 논문집 Vol.11 No.0201

In this paper, a comparison of classification ability between BPN( Back Propagation Neural Network) and k-NN (k-Nearest Neighbor) is made using the classification methods with patellar subluxation images and voice data. First, automatic diagnosis possibility of patellar subluxation symptoms, from CT images, is examined. Experimentally, BPN(a three-layered back propagation neural network) method is compared with k-NN(k nearest neighbor) method. A typical example of the conventional nonparametric recognition method is used, in the classification ability in order to evaluate the BPN method. Continuously, the same comparison in voice data is made as mentioned above. From the result of image recognition of patellar subluxation by using BPN method, Automatic diagnosis possibility is suggested in some degree. BPN method showed 9.2 percent higher than that of the k-NN classification method in the average recognition rate. Although k-NN classification is simple in theory, classification time was fairly long. Therefore, it seems that real time recognition is difficult. On the other hand, the BPN method is slow in learning time but is very fast in recognition time. If the number of dimensions of the samples is large, it can be said that BPN is better than k-NN in classification ability.

粉末 燻乾정어리 加工 및 貯藏中에 品質變化

金洙賢,李應昊,韓鳳浩,金世權 釜山水産大學校 1978 釜山水産大學 硏究報告 Vol.18 No.1

1976年 以來 우리나라에서 漁獲量이 急增하고 있는 정어리를 보다 效率的으로 食用化하기 위한 方法을 檢討하기 위하여 粉末燻乾정어리의 加工 및 貯藏에 관한 實驗을 하였다. 製品은 水分이 11.7%, 脂肪이 22.9%, 蛋白質은 61.7%였고, 原料 fillet에 대한 收率은 16.8%였으며, 담백한 맛과 燻乾品特有의 香氣가 있었다. 그리고 製品은 탄산가스 중에 貯藏하면 貯藏 중 酸化酸量의 變化가 거의 없고 品質을 安全하게 保藏할 수 있다는 結論을 얻었다. Since 1976 the catches of sardine have been increased rapidly in Korea. A processing method of smoked sardine meat powder and its quality changes during storage were studied. The fillets of raw sardine meat were boiled at 98℃ for 1 hour in a wire basket. After cooling at room temperature, the boiled fillets which small bones were removed were smoked at 60-70℃ for 8 hours. The second smoking was carried out under same conditions after 24 hours. The products were cleaned with brush and then powdered. The contents of moisture, lipid and protein were 11.7%, 22.9% and 61.7%, respectively. The yield of the product to the weight of raw fillet was 16.8%. According to the results of panel test, the product was well accepted having meaty taste and characteristic smoky flavor. The amount of oxidized acid in the product increased only a little during the storage in carbon dioxide gas.

MRI 아티팩트의 제거에 관한 연구

권진이,김응규 한밭대학교 정보통신전문대학원 2003 정보통신전문대학원 논문집 Vol.1 No.1

In this study, we present, an algorithm for canceling MRI artifacts due to some translational motion in patient’s head part. Assuming that the head part moves up and down due to patient's breathing which often makes problems during MRI imaging, rigid translational motions are treated here. Unlike the conventional iterative phase retrieval algorithm in which there is no guarantee for the convergence, this method is based on the MRI imaging process and the analysis of image property. A new constraint condition with which the image component and the motion component in the MRI signal can be separated by a simple algebraic operation is extracted After the x(read out) directional Fourier transformation of MRI signal is taken the y(phase encoding) directional spectrum phasing value is just an algebraic sum of the image component and the motion component. Meanwhile, at it is known that the density of subcutaneous fat area is almost uniform in the head tomographs, the density distribution along an y directional line on this fat area is regarded as symmetric shape. If the density function is symmetric, the phase of spectrum changes linearly with the position. Hence, the departure component methods, the motions in the x direction and the y direction are treated simultaneously. However, the x and the y directional motion is canceled by each different algorithm in two steps because the features of x and y directional motions are different from each other in this method. The x directional motion is estimated by detecting the edge boundary between the non-zero area and the zero area of the spectrum and is canceled by shifting the spectrum in inverse direction. The y directional motion is canceled by using the constraint condition which separates the true image component and the motion component. Finally, the effectiveness of this algorithm is shown by using a phantom image with simulated motion.

레이져스캐너 3차원계측에 있어서 다방향 데이터의 통합

金應珪,權寧島 大田産業大學校 1995 한밭대학교 논문집 Vol.12 No.0201

When objects are scanned satially by a laser-beam and mechanical mirror scanners, a technique is presentd for obtainment of three-dimensional data, and then measurement errors are often a significant problem. In this paper, first, a model of the systematic errors for a two-dimensional mirror scanner is developed based on the geometry of the laser scanning system. Parameter values are modified to reduce the different between measured and calculated values. This show that it is possible to measure three-dimensional data in a three-dimensional measurement system with enough precision for practical use. Futhermore, complete object data can be obtained without occlusion by combine scanner data from multiple directions.

중복 자궁 경부, 맹낭성 종질중격 및 동측 신장발생부전을 가진 교통성 중복자궁 1례

김윤진,전명권,이응수,최형민 인제대학교 백병원 2002 仁濟醫學 Vol.23 No.4

The class of uterine malformation known as communicating uteri is characterized by two uterocervical cavities connected by an isthmic communication. We experienced case of communicating double uterus with double cervix, blind hemivagina and ipsilateral renal agenesis. The complaint of the patient was odorous vaginal discharge with lower abdominal pain unrelated with menstrual cycle. We confirmed these Mullerian anomalies by transvaginal ultrasound. IVP(intravenous pyelography), MRI(magnetic resonance imaging), hysterosalpingography and hysteroscopy. After vaginal septoplasty, her complaint was subsided. We report this case with brief review of literatures.

참치 뼈를 이용한 Hydroxyapatite 세라믹 복합체의 합성 및 생체 친화성(제3보) : 인공체액에서의 Hydroxyapatite 세라믹 복합체간의 결합의 전자현미경 관찰 SEM Photographs of Bonding Properties between Hydroxyapatite Ceramics Composites in the Simulated Body Fluid

김세권,최진삼,이창국,변희국,전유진,이응호,박인용 대전산업대학교 반도체기술연구소 1999 半導體技術硏究所報 Vol.1 No.-

참치 뼈에서 추출한 hydroxyapatite를 출발물질로 여러 가지 세라믹 복합체를 제조하여 인공체액에서의 hydroxyapatite 복합체간의 화학결합을 조사하였다. hydroxyapatite 복합체들은 인공체액에서 4주 후부터 화학 결합성질을 나타내었다. 결합강도는 bioglass가 가장 강하게 나타났으며, 조성에 따른 의존성은 관찰할 수 없었다. 인공체액에서 복합체들은 이들의 경계면에서 불균일 핵생성 및 성장에 의해 화학적 결합으로 이루어졌다. Chemical bonding was investigated in the simulated body fluid of several selected hydroxyapatite-containing composites. The hydroxyapatite-containing composites chemically bonded with each other in the simulated body fluid after 4 weeks. Bioglass was strongly bonded in the simulated body fluid, but bonding strength was not depended on composition. Their composite bodies were chemically bonded by heterogeneous nucleation and growth at the interfaces of the specimens in the simulated body fluid.

주사형 정전용량 현미경의 개발

金應珪,權寧島 大田産業大學校 1996 한밭대학교 논문집 Vol.13 No.0201

In this study, a scanning capacitance microscopy(SCaM) by stage driving is proposed and presented some of the experimental results. SCaM is a microscopy which scans a surface of materials mechanically in two or two point five dimensions by a capacitance probe with a few tenth ㎛ size tip, and display images of the surface shape or capacitive distribution. The present target of the SCaM is 0.1㎛ resolution power which exceeds that of optical microscopy. This will become a powerful tool for inspecting ULSI pattern etched by X-ray, biological data, etc. The experimental system is composed based on a VHD video disk which captures the capacitance changes of the video disk surface and converts it into video signal.

웨이블릿 변환에 의한 정보 압축

김응규,권영도 한밭대학교 정보통신전문대학원 2005 정보통신전문대학원 논문집 Vol.3 No.1

본 연구에서는 웨이블릿 변환법에 기반한 정보압축을 제안한다. 신호와 화상의 압축 원리는 그 전력이 낮은 영역쪽으로 집중한다는 것을 이용한 비트의 할당, 즉 최적 양자화로, 주파수 분해방법으로써 이산코사인 변환법과 하부밴드 부호화 등이 실용화되고 있다. 본 연구에서 취급하는 웨이블릿 변환법도 그중의 한가지로 신호를 다중해상도로 분해해서 최적양자화를 행한다. 웨이블릿 변환은 주파수분할 방법 중의 하나로서 이산코사인 변환 또는 서브밴드 필터링 그리고 필터뱅크로 구현할 수 있다. 또한 공간적으로 국한(집중)되는 기저함수의 우수한 성질로부터 다른 방법에서 지적되어진 단점을 억제하는 효과가 있어, 다른 방법과 비교하면서 웨이블릿 변환의 장점에 대해서도 기술한다. In this study, on information compression based on the wavelet technique is described. The principle of signal or image compression is performed by optimization of quantization, that is the bit allocation taking advantage of their energy concentration in low frequency components. The wavelet transform is one of frequency decomposition, such as the discrete cosine transform or sub-band filtering, and it is also implemented as a filter bank. Wavelet transform with use of spatially localized basis function can reduce several drawbacks in conventional methods. The benefit of wavelet based compression method is described as comparing the transform method to another ones.

韓國産 달팽이의 化學成分에 관한 硏究 : (1) 민달팽이와 달팽이의 無機質 및 아미노酸組成 (1) Mineral Content and Amino Acid Composition of Korean Slug and Korean Land Snail

金世權,李應昊,文聖勳,姜玉珠 釜山水産大學校 1983 釜山水産大學 硏究報告 Vol.23 No.2

민달팽이 및 달팽이의 化學成分을 分析하여 營養學的 價値 및 漢方效果를 究明하기 위한 基礎資料를 얻고자 一般成分, 無機質, 構成아미노酸 및 遊離아미노酸을 分析한 結果는 다음과 같다. 민달팽이 및 달팽이의 蛋白質含量은 各各 11.6%, 14.7%였으며 灰分含量은 2.7% 및 2.1%이었다. 민달팽이의 無機物 중 K이 213.0㎎%로 가장 含量이 많았고, Na, P 및 Ca의 含量은 各各 142.9㎎%, 138.6㎎% 및 129.5㎎% 였다. Cu, Fe, Zn 및 Mn은 微量元素로서 檢出되었지만 Fe含量은 29.63㎎%로 매우 높은 含量이었다. 달팽이의 경우도 K이 288.2㎎%로 가장 그 含量이 많았고, 다음은 P이 173.7㎎%, Ca은 97.2㎎%, Na은 51.4㎎%, Mg은 30.8㎎%이었다. 微量元素 중 Zn 및 Fe含量은 各各 16.58㎎%, 8.64㎎%이었다. 構成아미노酸은 乾物重量으로서 민달팽이의 40.9%였고, glutamic acid, aspartic acid, leucine, arginine, lysine 및 glycine이 全體아미노산의 53.4%로 함량이 많았으며, cystine, methionine 및 hictidine은 그 含量이 적었다. 必須아미노酸 含量은 總아미노산의 33.0%를 차지하였으며, 이와같은 傾向은 달팽이의 경우도 비숫한 傾向이었다. 민달팽이의 遊離아미노酸 중 含量이 많은 것은 arginine, glutamic acid, alanine 및 serine이었고, 이들이 全體 아미노酸의 54.8%이었으며 반면에 methionine, tyrosine은 含量이 적었으며, 全體 遊離아미노酸 含量은 711.3㎎/100g이었다. 달팽이의 遊離아미노酸 중 含量이 많은 것은 alanine, glutamic acid, arginine, lysine이었으며 이들이 全體 아미노산의 53.6%를 차지한 반면 mthionine, tyrosine 및 cystine은 매우 낮은 含量이었으며 全體 遊離아미노酸含量은 1052.2㎎/100g이었다. The proximate compositions, mineral contents and amino acid compositions of Korean slug, Incilaria frusferi, and Korean land snail, Acusta despecta(GRAY), were determined. The contents of protein, lipid and ash in Korean slug were 11.6%, 2.8% and 2.7%, and those from Korean land snail were 14.7%, 2.2% and 2.1%, respectively. Among the minerals in Korean slug, potassium, sodium and calcium were 213.0 mg%, 142.9 mg% and 129.5 mg%, and those from Korean land snail were 288.2 mg%, 51.4 mg% and 97.2mg%, respectively. In the case of trace element in Korean slug, iron, manganese and copper were 29.63 mg%, 11.21 mg% and 1.64 mg%, and those from Korean land snail were 6.84 mg%, 6.85 mg%, and 2.74 mg%, respectively. The protein of Korean slug was composed of glutamic acid and aspartic acid as a major amino acids and considerable amount of leucine, arginine, lysine and glycine. The amount of essential amino acids except tryptophan was 40.5% of total amino acid. The amino acid composition of Korean land snail was similar to those of Korean slug. Among the free amino acid in Korean slug, arginine, glutamic acid, alanine and serine showed high quantity and the sum of these amino acids occupied 54.8% of total free amino acid. While methionine and tyrosine were poor in content. In the Korean land snail, alanine, glutamic acid, arginine and lysine were abundant amino acids, and the sum of these amino acids occupied 53.6% of total free amino acid. The quantities of methionine, tyrosine and cystine were, in contrast, very small.

신경회로망에 의한 2차원초음파의 화상재구성

金應珪,權寧島 大田産業大學校 1994 한밭대학교 논문집 Vol.11 No.0201

Comparing with light, Utrasonic has so small degenerating characteristic in solid and liquid matters that the measurement to use ultrasonic is widely used in medicine and biology. In this paper, echo process of ultrasonic transducer which is replace in circle type is done in the neural network and the method for image reconstruction of two dimensional target is presened. The image reconstruction in the same target its learned pattern and in unlearned target is done in this experiment. In result, although image reconstruction in unlearned target was not enough obtained, in the case of the same target as learned pattern, most of target images are reconstructed by this method and it is possible that two dimensional image is also reconstructed. It is hinted that three dimensional image reconstruction is possible in some target by improving the neural network structure and increasing the kinds and the number of tearing patterns.