• Proceedings of the Korean Society of Computer Information Conference
• /
• 2016.07a
• /
• pp.17-20
• /
• 2016

딥러닝 기법이 발달함에 따라 텍스트에 내재된 의미 및 구문을 어떠한 벡터 공간 상에 표현하기 위한 언어 모델이 활발히 연구되어 왔다. 이를 통해 자연어 처리를 기반으로 하는 감성 분석 및 문서 분류, 기계 번역 등의 분야가 진보되었다. 그러나 대부분의 언어 모델들은 텍스트에 나타나는 단어들의 일반적인 패턴을 학습하는 것을 기반으로 하기 때문에, 문서 요약이나 스토리텔링, 의역된 문장 판별 등과 같이 보다 고도화된 자연어의 이해를 필요로 하는 연구들의 경우 주어진 텍스트의 주제 및 의미를 고려하기에 한계점이 있다. 이와 같은 한계점을 고려하기 위하여, 본 연구에서는 기존의 LSTM 모델을 변형하여 문서 주제와 해당 주제에서 단어가 가지는 문맥적인 의미를 단어 벡터 표현에 반영할 수 있는 새로운 언어 학습 모델을 제안하고, 본 제안 모델이 문서의 주제를 고려하여 문장을 자동으로 생성할 수 있음을 보이고자 한다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2012.06b
• /
• pp.153-154
• /
• 2012

바디 센서 시스템 환경이란 사용자가 서기, 걷기, 뛰기 등의 행위를 통해 주기적으로 상황이 변하는 동적 환경이다. 이와 같은 시스템에서는 크기가 작고 저전력을 요구하는 센서가 탑재되기 때문에 효율적인 알고리즘을 적용하는 것은 매우 중요한 일이다. 모델체커는 최근 소프트웨어 모델 (Model)을 검증하는 도구로써 주어진 모델과 속성값을 통해 해당 모델의 검증 (Verification) 결과가 참인지 거짓인지 판별해 준다. 본 논문에서는 효율적인 바디 센서 시스템 구축을 위해 서기, 걷기, 뛰기라는 환경에서 개별적으로 동작되는 알고리즘을 모델링 하고 LTL(Linear Temporal Logic) 로 속성을 명세하여 NuSMV 모델 체커를 통해 해당 모델의 Safety와 Liveness를 검증한다.

• Journal of KIISE:Software and Applications
• /
• v.37 no.8
• /
• pp.599-610
• /
• 2010

Color laser printers are nowadays abused to print or forge official documents and bills. Identifying color laser printers will be a step for media forensics. This paper presents a new method to identify color laser printers with printed color images. Since different printer companies use their own printing process, each of printed papers from different printers has a little different invisible noise. After the wiener-filter is used to analyze the invisible noises from each printer, we extract some features from these noises by calculating a gray level co-occurrence matrix. Then, these features are applied to train and classify the support vector machine for identifying the color laser printer. In the experiment, we use total 2,597 images from 7 color laser printers. The results prove that the presented identification method performs well using the noise features of color printed images.

• The Journal of the Acoustical Society of Korea
• /
• v.14 no.1
• /
• pp.31-39
• /
• 1995

In this paper, we have constructed phoneme discriminant filters [PDF] according to the linear discriminant function. These discriminant filters do not follow the heuristic rules by the experts but the mathematical methods in iterative learning. Proposed system. is based on the piecewise linear classifier and error correction learning method. The segmentation of speech and the classification of phoneme are carried out simutaneously by the PDF. Because each of them operates independently, some speech intervals may have multiple outputs. Therefore, we introduce the unified coefficients by the output unification process. But sometimes the output has a region which shows no response, or insensitive. So we propose time windows and median filters to remove such problems. We have trained this system with the 549 monosyllables uttered 3 times by 3 male speakers. After we detect the endpoint of speech signal using threshold value and zero crossing rate, the vowels and consonants are separated by the PDF, and then selected phoneme passes through the following PDF. Finally this system unifies the outputs for competitive region or insensitive area using time window and median filter.

• The Journal of the Convergence on Culture Technology
• /
• v.8 no.4
• /
• pp.355-360
• /
• 2022

As the increase cost of domestic red pepper and the increase of imported red pepper, damage cases such as false labeling of the origin of red pepper powder are issued. Accordingly we need to determine quickly and accurately for the origin of red pepper powder. The used method for presently determining the origin has the limitation in that it requires a lot of cost and time by experimentally comparing and analyzing the components of red pepper powder. To resolve the issues, this study proposes machine learning algorithm to classifiy domestic and imported red pepper powder. We have built machine learning model with 53 components contained in red pepper powder and validated. Through the proposed model, it was possible to identify which ingredients are importantly used in determining the origin. In the near future, it is expected that the cost of determining the origin can be further reduced by expanding to various foods as well as red pepper powder.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.43 no.6
• /
• pp.851-863
• /
• 2023

The mandatory use of BIM(Building Information Model) in larger Korean public construction projects necessitates participants to have a comprehensive understanding of the relevant procedures and technologies, especially during the bidding stage. However, most small and medium-sized construction and engineering companies possess limited BIM proficiency and understanding. This hampers their ability to recognize bidding requirements and make informed decisions. To address this challenge, our study introduces a method to gauge the complexity of BIM requirements in bidding documents. This is achieved by integrating a morphological analyzer, which encompasses BIM bidding terminology, with a deep learning model. We investigated the effects of the parameters in our proposed deep learning model and examined its predictive validity. The results revealed an F1-score of 0.83 for the test data, indicating that the model's predictions align closely with the actual BIM performance challenges.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2019.10a
• /
• pp.861-863
• /
• 2019

사람 얼굴 생성을 목적으로 하는 Generative Adversarial Network(GAN)에서 판별자(discriminator)의 각 레이어에 대한 스펙트럴 정규화(spectral normalization) 적용에 따른 출력 이미지의 결과를 비교하였다. 또한 생성자(generator)에 적응 인스턴스 정규화(Adaptive Instance Normalization) 모듈의 삽입에 따른 출력 이미지의 결과를 기존 모델과 비교하고 분석하였다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2009.11a
• /
• pp.151-152
• /
• 2009

Grid 컴퓨팅 환경에서 Risk-resilient 한 Job 수행을 보장하기 위해 그 동안 Job migration 기법이 연구되어 왔으나, 자원 재선정 및 Job 이동/재할당에 따른 기준의 단순성으로 인해, Migration에 따른 Job 수행의 이득과 손실이 정확하게 판별되지 못한 경향이 있었다. 따라서 본고에서는 Job failure Rate을 바탕으로 특정 Job의 확률적 수행 지연 시간을 추정하고, 이를 이용하여 Migration gain을 평가하는 모델을 제안한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.479-480
• /
• 2023

생성형 인공지능은 학습의 기준을 파악하기 어려운 모델이다. 그 중 DCGAN을 분석하여 판별자를 통해 생성자의 학습 기준을 판단할 수 있는 하나의 방법을 제안하고자 한다. 그 과정에서 XAI 기법인 Grad-CAM을 활용하여 학습 시에 모델이 중요시하는 부분을 분석하여 적합한 학습과 학습에 적합하지 않은 데이터를 분석하는 방법을 소개하고자 한다.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.23 no.4
• /
• pp.340-348
• /
• 2021

Continuous monitoring of RGB (Red, Green, Blue) vegetation indices is important to apply remote sensing technology for the estimation of crop growth. In this study, we evaluated the performance of eight vegetation indices derived from soybean RGB images with various agronomic parameters under drought stress condition. Drought stress influenced the behavior of various RGB vegetation indices related soybean canopy architecture and leaf color. In particular, reported vegetation indices such as ExGR (Excessive green index minus excess red index), Ipca (Principal Component Analysis Index), NGRDI (Normalized Green Red Difference Index), VARI (Visible Atmospherically Resistance Index), SAVI (Soil Adjusted Vegetation Index) were effective tools in obtaining canopy coverage and leaf chlorophyll content in soybean field. In addition, the RGB vegetation indices related to leaf color responded more sensitively to drought stress than those related to canopy coverage. The PLS-DA (Partial Squares-Discriminant Analysis) results showed that the separation of RGB vegetation indices was distinct by drought stress. The results, yet preliminary, display the potential of applying vegetation indices based on RGB images as a tool for monitoring crop environmental stress.