• 대한방사선기술학회지:방사선기술과학
• /
• 제30권4호
• /
• pp.419-426
• /
• 2007

본 논문에서는 맛 인식을 위한 입력패턴벡터를 추출하고 패턴인식을 위한 맛(쓴맛, 단맛, 신맛, 짠맛)학습 알고리즘을 설계하였다. 입력패턴벡터의 구성을 위해 맛 활성화 신호의 세기가 사용되었고, 맛 패턴인식을 위한 알고리즘은 초기 참조벡터의 학습을 위해 SOM을 이용하였고, 종속 부류층의 출력뉴런의 부류지정을 위하여 out-star 학습법을 사용하였다. 제안된 알고리즘의 입력 층과 종속 클래스 층 사이의 연결강도는 SOM과 LVQ 알고리즘을 이용하여 초기 참조벡터의 설정 및 학습이 가능하게 하였다. 패턴벡터는 종속 부류층의 뉴런에 의해 종속 클래스로 분류하고, 종속 클래스 층과 출력 층 사이의 연결강도는 분류된 종속 부류를 클래스로 지정하는 학습을 하게 하였다. 패턴 분류를 위하여 제안된 학습알고리즘을 이용하여 시뮬레이션 되었고 기존의 LVQ 학습방식보다 우수한 분류성공률을 확인하였다.

• 한국컴퓨터정보학회논문지
• /
• 제11권5호
• /
• pp.95-103
• /
• 2006

본 논문에서는 뇌파인식을 위한 입력패턴벡터를 추출하고 패턴인식을 위한 뇌파 학습 알고리즘을 설계하였다. 입력패턴벡터의 구성을 위해 알파리듬과 베타리듬의 주파수와 진폭이 사용되었고, 뇌파패턴인식을 위한 알고리즘은 초기 참조벡터의 학습을 위해 SOM을 이용하고, 종속 부류층의 출력뉴런의 부류지정을 위하여 out-star 학습법을 사용하였다. 제안된 알고리즘의 입력 층과 종속 클래스 층 사이의 연결강도는 SOM과 LVQ 알고리즘을 이용하여 초기 참조벡터의 설정 및 학습이 가능하게 하였고, 패턴벡터를 종속 부류층의 뉴런에 의해 종속 클래스로 분류하고, 종속 클래스 층과 출력 층 사이의 연결강도는 분류된 종속 부류를 클래스로 지정하는 학습을 하게 된다. 뇌파 패턴 분류를 위하여 제안된 학습알고리즘을 이용하여 시뮬레이션 되었고 기존의 LVQ 학습방식보다 우수한 분류성공률을 확인하였다.

• 대한원격탐사학회지
• /
• 제39권3호
• /
• pp.339-353
• /
• 2023

고해상도 위성영상의 제공이 증가함에 따라 위성영상의 위치정확도 향상이 요구되고 있다. 이를 위해 기복변위를 제거하고 인공지물의 정위가 수립된 정사영상 생성의 중요성이 높아지고 있다. 본 논문에서는 기존에 구축된 건물 높이 데이터베이스를 이용하여 원본 위성영상에서의 건물 옥상면과 건물포함영역을 자동으로 추출하였다. 이후 추출된 건물 옥상면을 정위치 편집하여 건물 정위 레이어(layer)를 생성하였다. 추출된 건물포함영역을 이용하여 위성영상에서 건물영역을 공백 처리하여 비건물 정위 레이어를 생성하였다. 이후, 실감정사 건물레이어와 실감정사 비건물레이어를 중첩하여 최종 정사영상을 제작하였다. 본 연구에서 제안한 방법은 KOMPSAT-3 및 KOMPSAT-3A 위성영상을 이용해 실험하였으며, 실험 결과를 수치지형도와 중첩하여 검증을 수행하였다. 실험결과 건물 정위 레이어는 0.4 m의 위치 오차를 가지는 것으로 나타났다. 제안 방법을 통해 도심지역에 대한 자동 실감정사영상 생성의 가능성을 확인하였다.

• 한국측량학회:학술대회논문집
• /
• 한국측량학회 2009년도 춘계학술발표회 논문집
• /
• pp.11-13
• /
• 2009

본 연구는 최신의 수치지도 확보를 위한 갱신방안으로서 건축 및 다양한 SOC 건설공사에서 사용되고 있는 CAD 자료의 준공도면을 활용하여 수치지도를 갱신하기 위해 GIS 데이터로 변환하는 기법을 제시하였다. 변환과정은 레이어 추출, 객체변환, 좌표변환, 포맷변환으로 구성되며 CAD 데이터로부터 각 과정을 거쳐 변환된 GIS 데이터는 수치지도를 갱신할 수 있다.

• Toxicological Research
• /
• 제13권1_2호
• /
• pp.161-165
• /
• 1997

This study was conducted to investigate the skin irritation by transdermal administration of the three dyes. These dyes were originated from onion by using extraction method. By the order of extraction from onion, A-dye was obtained from onion by using water at 90-100$\circ$C. B-dye was extracted from A-dye with ethylacetate. After ethylacetate extraction from A-dye, the lower layer named as C-dye. Twenty-four New Zealand white rabbits were divided into three groups. The each groups was consisted of two subgroups according to high dose (extracted dyes) and low dose (the 100-fold dilutions of A-, Band C-dye). In primary skin irritation test of male New Zealand White rabbits, body temperature and weights were not significantly changed and blood cells were positioned in normal blood cell ranges of health rabbits. Primary irritation index was "0" in the test and control sites of all animals used in this study. By the results obtained in the present test, all dyes were evaluated as a non-irritant on the basis of the criteria of Draize.of Draize.

• Bulletin of the Korean Chemical Society
• /
• 제35권1호
• /
• pp.45-50
• /
• 2014

A fully packed capillary electrochromatographic (CEC) microchip with an on-column micro-solid phase extraction (SPE) bed for the preconcentration and separation of organic analytes was prepared. A linear microchannel with monodisperse colloidal silica packing was formed on a cyclic olefinic copolymer microchip with two reservoirs on both ends. Silver-cemented silica packing frit structure was formed at the entrance of the microchannel by electroless plating treatment as a base layer. A gold coating was formed on it by reducing $Au^{3+}$ to gold with hydroxylamine. Finally micro-SPE bed was formed by self-assembly adsorption of 1-dodecanethiol on it. Micro-SPE beds were about 100-150 ${\mu}m$ long. Approximately $10^3$ fold sensitivity enhancements for Sulforhodamine B, and Fluorescein in nM concentration levels were possible with 80 s preconcentration. Basic extraction characteristics were studied.

• Journal of Dairy Science and Biotechnology
• /
• 제17권1호
• /
• pp.11-15
• /
• 1999

The objective of this study was to extract and purity the antibacterial agent from the fermented milk with Lactobacillus helveticus CH-1. The extraction and purification of antibacterial agent from the Lb. helveticus fermented milk were carried out by methanol extraction, acetone extraction, Sephadex G-200 gel filteration and thin layer chromatography and the results were as followings. The antibacterial activity of methanol-acetone extraction showed antibacterial activity against test organisms, B. subtilis, E. coli, Pseu. fluorescens, Sal. typhimurium, Shi. flexneri, and Sta. aureus. Sephadex G-200 gel chromatography showed only antibacterial activity from 33 to 37th fractions of 60 fractions. The agent purified from TLC plate confirmed the antibacterial activity by the means of bioautography.

• Journal of Ginseng Research
• /
• 제45권1호
• /
• pp.1-21
• /
• 2021

Panax species have gained numerous attentions because of their various biological effects on cardiovascular, kidney, reproductive diseases known for a long time. Recently, advanced analytical methods including thin layer chromatography, high-performance thin layer chromatography, gas chromatography, high-performance liquid chromatography, ultra-high performance liquid chromatography with tandem ultraviolet, diode array detector, evaporative light scattering detector, and mass detector, two-dimensional high-performance liquid chromatography, high speed counter-current chromatography, high speed centrifugal partition chromatography, micellar electrokinetic chromatography, high-performance anion-exchange chromatography, ambient ionization mass spectrometry, molecularly imprinted polymer, enzyme immunoassay, 1H-NMR, and infrared spectroscopy have been used to identify and evaluate chemical constituents in Panax species. Moreover, Soxhlet extraction, heat reflux extraction, ultrasonic extraction, solid phase extraction, microwave-assisted extraction, pressurized liquid extraction, enzyme-assisted extraction, acceleration solvent extraction, matrix solid phase dispersion extraction, and pulsed electric field are discussed. In this review, a total of 219 articles published from 1980 to 2018 are investigated. Panax species including P. notoginseng, P. quinquefolius, sand P. ginseng in the raw and processed forms from different parts, geographical origins, and growing times are studied. Furthermore, the potential biomarkers are screened through the previous articles. It is expected that the review can provide a fundamental for further studies.

• Current Photovoltaic Research
• /
• 제1권1호
• /
• pp.59-62
• /
• 2013

We investigated the effects of ZnO thin film deposition methods on the performance of inverted polymer solar cells with a structure of ITO/ZnO/P3HT:PCBM/MoO3/Al. The ZnO thin films were deposited by various methods (spin coating of nanoparticles, sol-gel process, atomic layer deposition) and their morphology was analyzed by atomic force microscopy (AFM). The device with ZnO nanoparticle thin films showed the highest power conversion efficiency of 3 % with low series resistance and high shunt resistance. The superior performance of the device with the ZnO nanoparticle layer is attributed to better electron extraction capability.

• Journal of Multimedia Information System
• /
• 제9권3호
• /
• pp.177-182
• /
• 2022

A deep neural network (DNN) includes variables whose values keep on changing with the training process until it reaches the final point of convergence. These variables are the co-efficient of a polynomial expression to relate to the feature extraction process. In general, DNNs work in multiple 'dimensions' depending upon the number of channels and batches accounted for training. However, after the execution of feature extraction and before entering the SoftMax or other classifier, there is a conversion of features from multiple N-dimensions to a single vector form, where 'N' represents the number of activation channels. This usually happens in a Fully connected layer (FCL) or a dense layer. This reduced 2D feature is the subject of study for our analysis. For this, we have used the FCL, so the trained weights of this FCL will be used for the weight-class correlation analysis. The popular DNN models selected for our study are ResNet-101, VGG-19, and GoogleNet. These models' weights are directly used for fine-tuning (with all trained weights initially transferred) and scratch trained (with no weights transferred). Then the comparison is done by plotting the graph of feature distribution and the final FCL weights.