• International Journal of Computer Science & Network Security
• /
• 제23권1호
• /
• pp.112-119
• /
• 2023

Traditionally used for networking computers and communications, the Internet has been evolving from the beginning. Internet is the backbone for many things on the web including social media. The concept of social networking which started in the early 1990s has also been growing with the internet. Social Networking Sites (SNSs) sprung and stayed back to an important element of internet usage mainly due to the services or provisions they allow on the web. Twitter and Facebook have become the primary means by which most individuals keep in touch with others and carry on substantive conversations. These sites allow the posting of photos, videos and support audio and video storage on the sites which can be shared amongst users. Although an attractive option, these provisions have also culminated in issues for these sites like posting offensive material. Though not always, users of SNSs have their share in promoting hate by their words or speeches which is difficult to be curtailed after being uploaded in the media. Hence, this article outlines a process for extracting user reviews from the Twitter corpus in order to identify instances of hate speech. Through the use of MPCA (Modified Principal Component Analysis) and ECNN, we are able to identify instances of hate speech in the text (Enhanced Convolutional Neural Network). With the use of NLP, a fully autonomous system for assessing syntax and meaning can be established (NLP). There is a strong emphasis on pre-processing, feature extraction, and classification. Cleansing the text by removing extra spaces, punctuation, and stop words is what normalization is all about. In the process of extracting features, these features that have already been processed are used. During the feature extraction process, the MPCA algorithm is used. It takes a set of related features and pulls out the ones that tell us the most about the dataset we give itThe proposed categorization method is then put forth as a means of detecting instances of hate speech or abusive language. It is argued that ECNN is superior to other methods for identifying hateful content online. It can take in massive amounts of data and quickly return accurate results, especially for larger datasets. As a result, the proposed MPCA+ECNN algorithm improves not only the F-measure values, but also the accuracy, precision, and recall.

• 한국의상디자인학회지
• /
• 제14권1호
• /
• pp.81-96
• /
• 2012

The purpose of this study is to provide fundamental information for standardization of 3D body measurement. This research analyzes errors occurring in the process of extracting body size from 3D body scan data. First, as a result of analyzing basic state of the 3D body scanner's calibration, the point number of each section was almost the same, while the right and left as well as the front and back coordinates of the center of gravity are not, showing unstable data. Nevertheless, the latter does not influence on the size of cylinder such as width and circumference. Next, we analyzed point coordinates variations of scan data on a mannequin nude by life casting. The result was great deflection in case of complicated or horizontal sections including the reference point beyond proper distance from centers of four cameras. In case of the mannequin's size, accuracy proves comparatively high in that measurement errors in height, width, depth, and length dimension occurred all within allowable errors, only except chest depth, while there were a lot of measurement errors in a circumference dimension. Secondly, analysis of accuracy of automatic extraction identification program algorithm presented that a semi-automatic measurement program is better than an automatic measurement program. While both of them ate very acute in parts related to crotch, they are not in armpit related parts. Therefore, in extracting of human body size from 3D scan data, what really matters seems to parts related to armpits.

• 한국축산식품학회지
• /
• 제38권2호
• /
• pp.362-375
• /
• 2018

This study proposed a rapid microscopic examination method for pork freshness evaluation by using the self-assembled hyperspectral microscopic imaging (HMI) system with the help of feature extraction algorithm and pattern recognition methods. Pork samples were stored for different days ranging from 0 to 5 days and the freshness of samples was divided into three levels which were determined by total volatile basic nitrogen (TVB-N) content. Meanwhile, hyperspectral microscopic images of samples were acquired by HMI system and processed by the following steps for the further analysis. Firstly, characteristic hyperspectral microscopic images were extracted by using principal component analysis (PCA) and then texture features were selected based on the gray level co-occurrence matrix (GLCM). Next, features data were reduced dimensionality by fisher discriminant analysis (FDA) for further building classification model. Finally, compared with linear discriminant analysis (LDA) model and support vector machine (SVM) model, good back propagation artificial neural network (BP-ANN) model obtained the best freshness classification with a 100 % accuracy rating based on the extracted data. The results confirm that the fabricated HMI system combined with multivariate algorithms has ability to evaluate the fresh degree of pork accurately in the microscopic level, which plays an important role in animal food quality control.

• 분석과학
• /
• 제16권2호
• /
• pp.110-116
• /
• 2003

사용 후 핵연료의 화학특성 연구를 위하여 요오드의 분리와 정량에 관한 연구를 수행하였다. 사용 후 핵연료를 용해시키는 과정에서 핵연료 중에 CsI로 존재하는 요오드가 $I_2$로 산화되어 휘발되지 않도록 질산과 염산의 혼합산 (80:20 mol%)을 이용하여 비휘발성 ${IO_3}^-$­로 안정화시켰다. 2.5 M $HNO_3$ 매질에서 $NH_2OH{\cdot}HCl$을 이용하여 $I_2$로 환원시킨 후 사염화탄소로 추출하여 우라늄과 핵분열생성물로부터 분리, 회수하였다. 0.1 M $NaHSO_3$을 사용하여 요오드를 역추출하였으며 수용액층으로 회수된 요오드를 이온 크로마토그래피로 정량하였다. 방사성 물질 분석에 적합한 이온 크로마토그래피/차폐 시스템을 구성하였으며 42,000~44,000 MWd/MtU 의 연소도를 갖는 사용후핵연료를 대상으로 요오드를 분석한 결과 Origin 2 연소도 전산코드에 의한 계산결과인 $324.5{\sim}343.6{\mu}g/g$와는 -8.3~-0.5%의 편차를 나타내었다.

• 인지과학
• /
• 제9권1호
• /
• pp.1-11
• /
• 1998

본 논문에서는 인간의 정보처리 과정에 관한 심리학적 실험에 바탕을 두고 인간이 사용하고 있는 것으로 생각되는 특징을 이용하여 이를 문자 인식에 적용하였다. 인간의 경우 화소단위의 정보뿐만 아니라 일정지역의 정보를 함께 처리하는 경향이 있다. 그러므로 일정지역에 대한 정보를 표시하는 영역 특징을 정의하고 정의된 이 영역 특징과 기존의 화소단위 특징들을 결합하였다. 사용한 특징으로는 영역 특징에 기반 한 초등 적 분석결과, 영역특징을 포함한 망 특징, 교차거리와 특징 그리고 기울기 특징들이다. 성능 평가 실험은 필기 한글자모, 숫자 그리고 대소영문자를 대상으로 하였으며, 인식기는 역전과 학습 방법을 이용한 신경망 인식기를 사용하였다. 각각의 인식 결과는 90.27∼93.25%, 98.00% 그리고 79.73∼85.75였다. 영역 특징과 유사한 UDLRH 특징을 대상으로 비교한 결과 전체적으로 1∼2% 정도 인식률 향상이 있었으며 인간이 판단하기에 보다 납득하기 쉬운 오 인식 성향을 보였다.

• 대한전자공학회논문지TC
• /
• 제43권10호
• /
• pp.27-34
• /
• 2006

본 논문은 3[Cm] 이내의 Tag-to-Reader 잡음 공간에서도 기계 ID 인식을 정확하게 보장하는 초소형($1.4{\times}2.8[Cm^2]$) 13.56[MHz] RFID Reader를 구현하였다. 그 RFID 시스템의 작동을 위하여, 먼저, 본 논문은 13.56[MHz] RFID Air Interface ISO/IEC 규격을 따르는 전파 전파에서 후방 산란의 페이딩 모델과 Loop Antenna를 설계하였고, 다음으로 초소형 RFID RF 이슈들을 측정하고 분석하여 자동으로 경로 선택된 RF 스위칭 회로와 펌웨어의 작동 관계를 제안하였으며, 끝으로, 초소형 Reader의 본체로서 국제표준규격 ISO/IEC 18000-3이 정의한 13.56[MHz] RFID 신호의 반송과 동시에 $1{\sim}2$개 기계 ID 정보의 추출과 오류 예방을 위하여 제작된 DSP(Digital Signal Processor) 보드와 소프트웨어 기능을 제시하였다.

• 대한치과교정학회지
• /
• 제23권4호
• /
• pp.565-606
• /
• 1993

The purpose of this study was to evaluate the change of before and after treatment of anterior openbite malocclusions treated by Multiloop Edgewise Archwire technique. The openbite sample consisted of 4 male and 12 female adults, treated with nonextraction or third molar extraction. The normal sample consisted of 58 subjects, which have pleasing facial profile and normal occlusion and no experience of orthodontic or prosthodontic treatment. The 58 subjects of normal sample were subdivided by cephalemetric vertical relationship of face. The 40 subjects, cephalometric vertical relationship of face was in normal range, classified as Normal Sample group 1. The 18 subjects, increased cephalometric vertical relationship of face, classified as Normal Sample group 2. The computerized cephalometric analysis was accomplished with 50 reference points for 22 skeletal measurements, 46 dentoalveolar measurements, 8 soft tissue measurements. Statistical analysis of the data was carried out with paired t-test, Student's t-test, and DUNCAN test using SAS(PC version), The results were as follows : 1. There were no statistically significant differences in skeletal measurement between before and after treatment. The major changes were in dentoalveolar region. 2. After treatment, the long axis of maxillary and mandibular posterior teeth were distally tipped-back, and uprighted to bisected occlusal plane. The interincisal angle was increased. 3. There were no statistically significant increase in the upper posterior dental height and statistically significant decrease in the lower posterior dental height. The upper anterior dental height was increased, but there was no statistically significant increase in the absolute upper anterior dental hight. The lower anterior dental height was increased. 4. After treatment, the maxillary occlusal plane to palatal plane angle and the mandibular occlusal plane to mandibular plane angle were statistically significant increased. Then, there were no statistically significant difference between after treatment group and normal sample group 2. 5. After treatment, the percentage of upper lip length to upper anterior dental height was decreased. Then, There were no statistically significant difference between after treatment group and normal sample group 2.

• 전자공학회논문지CI
• /
• 제38권3호
• /
• pp.61-67
• /
• 2001

본 논문에서는 한국어 숫자음 인식을 위해 음성과 영상 정보를 사용하고, 음성에 사용하는 선형예측계수 알고리즘을 영상에 적용하는 방법을 제안한다. 입력으로 얻어지는 음성신호는 0.95의 매개변수를 통해 고역 신호가 강조되고, 해밍창과 자기상관 분석, Levinson-Durbin 알고리즘에 의해 13차 선형예측계수를 구한다. 마찬가지로, 그레이 영상신호도, 음성의 자기상관 분석, Levinson-Durbin 알고리즘을 사용하여 13차의 2차원 선형예측계수를 구한다, 이러한 음성/영상 신호에 대한 선형예측계수들은 다층 신경회로망에 적용하여 학습이 이루어졌고, 각 레벨의 잡음이 섞인 음성신호를 적용한 결과, 숫자음 '3', '5', '9' 에서 음성만으로 인식한 결과보다 훨씬 좋은 인식결과를 얻을 수 있었다. 결과적으로, 본 연구에서는 영상 신호의 2차원 선형 예측 계수들이 음성인식에 사용될 경우, 특징 추출에 따른 부가적인 알고리즘이 새로 고안될 필요가 없이, 음성특징 계수를 추출하는 방법을 그대로 사용할 수 있으며, 또한 데이터량과 인식율이 잡음 환경에서 보다 향상되는 효율적인 방법을 제시하고 있음을 알 수 있었다.

• 한국컴퓨터그래픽스학회논문지
• /
• 제23권5호
• /
• pp.29-38
• /
• 2017

본 논문은 나뭇가지 패턴을 의자의 골격에 임의로 합성하는(Tree-Inspired Chair) 모델링 기법을 제안한다. 여러 개의 입력모델을 합성하는 기존 기법과 다르게, 제안 기법은 하나의 메쉬만 사용하여, 사용자가 원하는 부분의 contour mesh로부터 나무 성장 시뮬레이션으로 생성된 패턴을 갖는 의자 모델링이 가능하다. 우리는 나뭇가지 패턴을 생성시킬 영역 contour mesh를 효율적으로 추출하기 위하여 새로운 기법을 제안한다. 우선, 입력된 모델의 face 면적에 기반한 contour mesh를 생성하고, 그 메쉬의 앞뒷면 정보를 이용하여 연결정보가 복원된 skeleton mesh를 생성한다. 또한, 입력 모델의 형상과 유사하게 나뭇가지 패턴을 생성하기 위해 형상 표면의 tangent vector를 고려하는 3-way 나무성장 시뮬레이션 기법을 제안한다. 제안기법은 기존의 가구 모델을 이용하여 간단한 파라미터의 조작만으로 나뭇가지 형상과 가구 모델의 골격을 결합하는 새로운 형태의 가구 모델링을 보여준다. 우리는 실험을 통하여 제안 기법의 성능과 유효성을 보여주었다.

• Nuclear Engineering and Technology
• /
• 제36권4호
• /
• pp.357-363
• /
• 2004

A prototype long pulse ion source was developed, and the beam extraction experiments of the ion source were carried out at the Neutral Beam Test Stand (NBTS) of the Korea Superconducting Tokamak Advanced Research (KSTAR). The ion source consists of a magnetic bucket plasma generator, with multi-pole cusp fields, and a set of tetrode accelerators with circular apertures. Design requirements for the ion source were a 120kV/65A deuterium beam and a 300 s pulse length. Arc discharges of the plasma generator were controlled by using the emission-limited mode, in turn controlled by the applied heating voltage of the cathode filaments. Stable and efficient arc plasmas with a maximum arc power of 100 kW were produced using the constant power mode operation of an arc power supply. A maximum ion density of $8.3{\times}10^{11}\;cm^{-3}$ was obtained by using electrostatic probes, and an optimum arc efficiency of 0.46 A/kW was estimated. The accelerating and decelerating voltages were applied repeatedly, using the re-triggering mode operation of the high voltage switches during a beam pulse, when beam disruptions occurred. The decelerating voltage was always applied prior to the accelerating voltage, to suppress effectively the back-streaming electrons produced at the time of an initial beam formation, by the pre-programmed fast-switch control system. A maximum beam power of 0.9 MW (i.e. $70\;kV{\times}12.5\;A$) with hydrogen was measured for a pulse duration of 0.8 s. Optimum beam perveance, deduced from the ratio of the gradient grid current to the total beam current, was $0.7\;{\mu}perv$. Stable beams for a long pulse duration of $5{\sim}10\;s$ were tested at low accelerating voltages.