• 한국항해학회지
• /
• 제24권4호
• /
• pp.227-234
• /
• 2000

In general, neural networks are widely used for the category classification of multispectral images. Since the input multispectral images into neural networks we, however, low contrast images, neural networks converge very slowly and are of bad performance. To overcome this problem, we propose a new image enhancement method which consists of smoothing process, finding the main valley and enhancement process. In addition the enhanced images by the proposed method are used as the input of neural networks for the category classification. When the new category classification method is applied to multispectral LANDSAT TM images, we verified that the neural networks converge very lastly and that the overall category classification performance is improved.

• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2021년도 춘계학술대회
• /
• pp.338-339
• /
• 2021

본 논문에서는 변환 학습을 기반으로 한 다중 클래스 이미지 장면 분류 방법을 제안한다. 이미지 분류를 위해 대형 이미지 데이터 세트 ImageNet에 대해 사전 학습 한 ResNet (ResNet) 모델을 사용하는 방법이다. CNN 모델의 이미지 분류 방법에 비해 분류 정확도 및 효율성을 크게 향상시킬 수 있다.

• 한국컴퓨터정보학회논문지
• /
• 제19권7호
• /
• pp.1-8
• /
• 2014

클러스터링은 대표적인 비교사 학습 방법의 하나로 균일한 특성을 가지는 데이터를 군집으로 묶기 위해 사용된다. 균일한 특성을 가지는 데이터 부분집합을 문맥으로 정의하고 문맥 내에서 국부적으로 분류를 행하는 융합 방법이 사용되고 있지만 클러스터링은 비교사 학습 방법이라는 한계로 인해 클러스터링 결과로 만들어지는 문맥이 분류에 있어 최선임을 보장하기 어렵다. 이 논문에서는 생성된 클러스터를 문맥으로 가정하고 각 문맥에서 분류를 시행하는 경우 최소의 오류를 보일 수 있는, 분류를 고려한 클러스터링 기법을 제안한다. 제안하는 방법은 선형 판별 분석에서와 유사하게 클러스터 내 동일한 클래스에 속하는 데이터 쌍은 작은 거리 값을, 서로 다른 클래스에 속하는 데이터 쌍은 큰 거리 값을 가지도록 하기 위한 제약 조건을 적용하여 분류 오류를 줄이도록 하였다. 제안한 방법의 실효성은 실험 결과를 통해 확인할 수 있다.

• 한국대기환경학회지
• /
• 제12권4호
• /
• pp.369-376
• /
• 1996

Major atmospheric stability classification methods were evaluated with meteorological data obtained by scoustic sounding profiler (SODAR/RASS) in Seoul. The Psequill classificatio method, the method most widely used because of its good agreement in respect of synoptic scope under the steady state, fails to describe the time lag, the response time on stability by heating or cooling caused by daily insolation or noctrunal surface radiation. Horizontal and vertical standard deviation of wind fluctuation $(\sigma_A and \sigma_E)$ method tend to classify night-time stable condition (E, F class) into unstable condition (A, B class). The classification matrix tables for Vogt's vertical temperature difference and wind speed using method ($\Delta$T $\cdot$ U) and bulk Richardson number (Rb) were amended for practical use over Seoul. The modified tables for $\Delta$T $\cdot$ U and Rb method were made by using comprehensive frequency distribution from Pasquill's method and other existing results, and the correlation coefficient(r) was equal to 0.829. It was confirmed that atmospheric stability could be changed with monitoring site characteristics, height and vertical difference between sensors of monitoring station, and classification method itself.

• 조명전기설비학회논문지
• /
• 제25권12호
• /
• pp.21-26
• /
• 2011

This is a study to establish regulations against light pollution for lighting on roads. Many kinds of light pollution by luminaire on roads was calculated and analyzed by applying the classification method of luminaires(Cut-off classification of IDA-IESNA, BUG Rating Classification) and the calculation method of Upward Lighting Ratio of CIE among measures to prevent light pollution that international lighting organizations suggest. As a result of the analysis, it was found that the regulation by Cutoff of IESNA and ULR classification of CIE could be one for scattered light of light pollution compared to BUG classification but is not sufficient for the regulation of light tresspass or glare. BUG classification by each lighting zone was suggested as threshold value of the light pollution regulation considering domestic conditions.

• 한국방송∙미디어공학회:학술대회논문집
• /
• 한국방송공학회 2009년도 IWAIT
• /
• pp.525-528
• /
• 2009

Recently, image classification has been an important task in various fields. Generally, the performance of image classification is not good without the adjustment of image features. Therefore, it is desired that the way of automatic feature extraction. In this paper, we propose an image classification method which adjusts image features automatically. We assume that texture features are useful in image classification tasks because natural images are composed of several types of texture. Thus, the classification accuracy rate is improved by using distribution of texture features. We obtain texture features by calculating image features from a current considering pixel and its neighborhood pixels. And we calculate image features from distribution of textures feature. Those image features are adjusted to image classification tasks using Genetic Algorithm. We apply proposed method to classifying images into "head" or "non-head" and "male" or "female".

• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2019년도 춘계학술대회
• /
• pp.155-156
• /
• 2019

Classification and Regression Tree (CART), SVM (Support Vector Machine) 및 k-nearest neighbor classification (kNN)과 같은 기존 기계 학습 기반 감정 분석 방법은 정확성이 떨어졌습니다. 본 논문에서는 개선 된 kNN 분류 방법을 제안한다. 개선 된 방법 및 데이터 정규화를 통해 정확성 향상의 목적이 달성됩니다. 그 후, 3 가지 분류 알고리즘과 개선 된 알고리즘을 실험 데이터에 기초하여 비교 하였다.

• Smart Structures and Systems
• /
• 제23권1호
• /
• pp.71-79
• /
• 2019

In general, it may be advantageous to measure the pressure pulsation near a valve to detect a valve defect in a linear compressor. However, the acceleration signals are more advantageous for rapid classification in a mass-production line. This paper deals with the performance improvement of fault classification using only the compressor-shell acceleration signal based on the relation between the refrigerant pressure pulsation and the shell acceleration of the compressor. A transfer function was estimated experimentally to take into account the signal noise ratio between the pressure pulsation of the refrigerant in the suction pipe and the shell acceleration. The shell acceleration signal of the compressor was modified using this transfer function to improve the defect classification performance. The defect classification of the modified signal was evaluated in the acceleration signal in the frequency domain using Fisher's discriminant ratio (FDR). The defect classification method was validated by experimental data. By using the method presented, the classification of valve defects can be performed rapidly and efficiently during mass production.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제18권2호
• /
• pp.420-437
• /
• 2024

There are some problems in network traffic classification (NTC), such as complicated statistical features and insufficient training samples, which may cause poor classification effect. A NTC architecture based on one-dimensional Convolutional Neural Network (CNN) and transfer learning is proposed to tackle these problems and improve the fine-grained classification performance. The key points of the proposed architecture include: (1) Model classification--by extracting normalized rate feature set from original data, plus existing statistical features to optimize the CNN NTC model. (2) To apply transfer learning in the classification to improve NTC performance. We collect two typical network flows data from Youku and YouTube, and verify the proposed method through extensive experiments. The results show that compared with existing methods, our method could improve the classification accuracy by around 3-5%for Youku, and by about 7 to 27% for YouTube.

• 응용통계연구
• /
• 제19권3호
• /
• pp.587-598
• /
• 2006

본 논문에서는 마이크로어레이 (microarray) 자료에 판별분석을 적용 시 나타나는 고차원 및 소표본 문제의 해결방법으로서 직교요인을 새로운 특징변수로 사용한 비모수적 국소선형 로지스틱 판별분석을 제안한다. 제안된 방법은 국소우도에 기반한 것으로서 다범주 판별분석에 적용될 수 있으며, 고려된 직교인자는 주성분 요인, 부분최소제곱 요인, 인자분석 요인 등이다. 대표적인 두 가지 실제 마이크로어레이 자료에 적용한 결과 직교요인들 중에서 부분최소제곱 요인을 특징변수로 사용한 경우 고전적인 통계적 판별분석보다 향상된 분류 능력을 나타내고 있음을 확인하였다.