• 한국정보처리학회논문지
• /
• 제2권6호
• /
• pp.995-1001
• /
• 1995

패턴인식은 전처리, 특징추출, 식별의 과정을 거쳐 인식을 하게된다. 식별과정 에서 여러개의 패턴이 흩어져 있을 경우에 유사한 패턴끼리 클러스터링을 위하여 한 카테고리 내에서 패턴을 분할하게 된다. 클러스터링 방법에는 통계적인 방법으로 k-means 방법, ISODATA알고리즘등이 있으며〔1〕, 최근에는 신경망에 의한 클러스터링 방법으로 T, Kohonen의 LVQ(Learning Vector Quantization)가 주로 이용되었다〔6〕. Nikhil R, Pal. et al은 LVQ알고리즘을 보다 개선한 방법으로 GLVQ(Generalized LVQ, 1993)를 제안하였다〔4〕.본 논문은 GLVQ 알고니즘으로 패턴을 클러스터링 할 경우에 효율적인 특징추출 방법을 제안한다. 본 논문에서는 20명의 필기체 숫자 0에서 9까지 의 200개 패턴을 여러 가지 방법으로 특징 추출하여 GLVQ알고리즘으로 10개(0-9의 패턴) 의 클러스터로 구분하고, 해당 클러스터에서 오분류되는 패턴의 비율로서 그 효율성을 비교 하였다. 그 결과 투영조합 방법을 이용하였을 경우 98.5%의 분류율을 나타내었다.

• International Journal of Contents
• /
• 제13권2호
• /
• pp.66-74
• /
• 2017

Patent classification is becoming more critical as patent filings have been increasing over the years. Despite comprehensive studies in the area, there remain several issues in classifying patents on IPC hierarchical levels. Not only structural complexity but also shortage of patents in the lower level of the hierarchy causes the decline in classification performance. Therefore, we propose a new method of classification based on different criteria that are categories defined by the domain's experts mentioned in trend analysis reports, i.e. Patent Landscape Report (PLR). Several experiments were conducted with the purpose of identifying type of features and weighting methods that lead to the best classification performance using Support Vector Machine (SVM). Two types of features (noun and noun phrases) and five different weighting schemes (TF-idf, TF-rf, TF-icf, TF-icf-based, and TF-idcef-based) were experimented on.

• The Journal of the Acoustical Society of Korea
• /
• 제23권3E호
• /
• pp.91-97
• /
• 2004

In this paper, we propose a classification system of TV program scenes based on audio information. The system classifies the video scene into six categories of commercials, basketball games, football games, news reports, weather forecasts and music videos. Two type of audio feature set are extracted from each audio frame-timbral features and coefficient domain features which result in 58-dimensional feature vector. In order to reduce the computational complexity of the system, 58-dimensional feature set is further optimized to yield l0-dimensional features through Sequential Forward Selection (SFS) method. This down-sized feature set is finally used to train and classify the given TV program scenes using κ -NN, Gaussian pattern matching algorithm. The classification result of 91.6% reported here shows the promising performance of the video scene classification based on the audio information. Finally, the system stability problem corresponding to different query length is investigated.

• 대한임베디드공학회논문지
• /
• 제16권1호
• /
• pp.17-27
• /
• 2021

In this paper, we develop a galaxy image classification method based on hand-crafted features and machine learning techniques. Additionally, we provide an empirical analysis to reveal which combination of the techniques is effective for galaxy image classification. To achieve this, we developed a framework which consists of four modules such as preprocessing, feature extraction, feature post-processing, and classification. Finally, we found that the best technique for galaxy image classification is a method to use a median filter, ORB vector features and a voting classifier based on RBF SVM, random forest and logistic regression. The final method is efficient so we believe that it is applicable to embedded environments.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2013년도 춘계학술발표대회
• /
• pp.204-207
• /
• 2013

Music genre classification is an essential component for music information retrieval system. There are two important components to be considered for better genre classification, which are audio feature extraction and classifier. This paper incorporates two different kinds of features for genre classification, timbral texture and rhythmic content features. Timbral texture contains several spectral and Mel-frequency Cepstral Coefficient (MFCC) features. Before choosing a timbral feature we explore which feature contributes less significant role on genre discrimination. This facilitates the reduction of feature dimension. For the timbral features up to the 4-th order central moments and the covariance components of mutual features are considered to improve the overall classification result. For the rhythmic content the features extracted from beat histogram are selected. In the paper Extreme Learning Machine (ELM) with bagging is used as classifier for classifying the genres. Based on the proposed feature sets and classifier, experiment is performed with well-known datasets: GTZAN databases with ten different music genres, respectively. The proposed method acquires the better classification accuracy than the existing approaches.

• 한국지능시스템학회논문지
• /
• 제15권4호
• /
• pp.492-497
• /
• 2005

영상 데이터와 같이 큰 차원을 가지는 입력 자료들을 분류하고자 할 경우, 입력 자료의 차원을 줄일 수 있는 특징을 추출하는 전처리 과정은 매우 중요하다. 특징 추출(feature extraction)을 위해 PCA, ICA, LDA, MLP 등의 다양한 기법들이 개발되었는데 이러한 기법들은 PCA, ICA와 같은 무감독 방식의 기법(unsupervised algorithm)과 LDA, MLP와 같은 감독 방식의 기법(supervised algorithm)으로 구분할 수 있다. 이 중에서, 감독 방식의 경우는 입력 정보와 함께 클래스 정보를 사용하기 때문에 데이터를 분류하기에 더 좋은 특징들을 뽑아낼 수 있다. 본 논문에서는 무감독 방식 기법인 PCA에 기반 하면서도 클래스 정보를 사용하여 자료 분류에 더욱 적합한 특징들을 추출할 수 있는 기법인 PCA-FX를 제안하였다. 제안한 기법에 의해 추출된 특징을 이용할 경우의 인식 성능을, Yale face database를 사용하여 다른 기법들의 성능과 비교하였다.

• 한국소음진동공학회:학술대회논문집
• /
• 한국소음진동공학회 2005년도 추계학술대회논문집
• /
• pp.127-130
• /
• 2005

There are many methods in feature extraction have been developed. Recently, principal components analysis (PCA) and independent components analysis (ICA) is introduced for doing feature extraction. PCA and ICA linearly transform the original input into new uncorrelated and independent features space respectively In this paper, the feasibility of using nonlinear feature extraction will be studied. This method will employ the PCA and ICA procedure and adopt the kernel trick to nonlinearly map the data into a feature space. The goal of this study is to seek effectively useful feature for faults classification.

• 한국측량학회:학술대회논문집
• /
• 한국측량학회 2007년도 춘계학술발표회 논문집
• /
• pp.183-186
• /
• 2007

Hyperspectral remote sensing data contain plenty of information about objects, which makes object classification more precise. Over the past several years, different algorithms for the classification of hyperspectral remote sensing images have been developed. In this study, we proposed method based on absorption band extraction and Gravity type model to solve hyperspectral image classification problem. In contrast to conventional methods that are based on correlation techniques, this method is simple and more effective. The proposed approach was tested to evaluate its effectiveness. The evaluation was done by comparing the results of preexiting SFF(Spectral Feature Fitting) classification method. The evaluation results showed the proposed approach has a good potential in the classification of hyperspectral images.

• 한국정보통신학회논문지
• /
• 제24권10호
• /
• pp.1341-1347
• /
• 2020

부정맥 분류를 위한 기존 연구들은 분류의 정확성을 높이기 위해 신경회로망(Artificial Neural Network), 기계학습(Machine Learning) 등을 이용한 방법이 연구되어 왔다. 특히 딥러닝은 신경회로망의 문제인 은닉층 개수의 한계를 해결함으로 인해 인공 지능 기반의 부정맥 분류에 많이 사용되고 있다. 본 연구에서는 AR 기반의 특징점 추출과 딥러닝을 통한 부정맥 분류 방법을 제안한다. 이를 위해 먼저 잡음을 제거한 ECG 신호에서 R파를 검출하고 자기 회귀 모델을 통하여 최적의 QRS와 RR간격을 추출하였다. 이후 딥러닝을 통한 지도학습 방법으로 가중치를 학습시키고 부정맥을 분류하였다. 제안된 방법의 타당성 평가를 위해 MIT-BIH 부정맥 데이터베이스를 통해 각 파라미터에 따른 훈련 및 분류 정확도를 확인하였다. 성능 평가 결과 PVC는 약 97% 이상의 평균 분류율을 나타내었다.

• 대한원격탐사학회지
• /
• 제20권5호
• /
• pp.315-327
• /
• 2004

Nowadays, interests in land cover classification using not only multi-sensor images but also thematic GIS information are increasing. Often, although useful GIS information for the classification is available, the traditional MLE (maximum likelihood estimation techniques) does not allow us to use the information, due to the fact that it cannot handle the GIS data properly. This paper propose two extended MLE algorithms that can integrate both remote sensing images and GIS thematic data for land-cover classification. They include modified MLE and Bayesian predictive likelihood estimation technique (BPLE) techniques that can handle both categorical GIS thematic data and remote sensing images in an integrated manner. The proposed algorithms were evaluated through supervised land-cover classification with Landsat ETM+ images and an existing land-use map in the Gongju area, Korea. As a result, the proposed method showed considerable improvements in classification accuracy, when compared with other multi-spectral classification techniques. The integration of remote sensing images and the land-use map showed that overall accuracy indicated an improvement in classification accuracy of 10.8% when using MLE, and 9.6% for the BPLE. The case study also showed that the proposed algorithms enable the extraction of the area with land-cover change. In conclusion, land cover classification results produced through the integration of various GIS spatial data and multi-spectral images, will be useful to involve complementary data to make more accurate decisions.