• International Journal of Fuzzy Logic and Intelligent Systems
• /
• 제14권4호
• /
• pp.332-339
• /
• 2014

Face classification has wide applications in security and surveillance. However, this technique presents various challenges caused by pose, illumination, and expression changes. Face recognition with long-distance images involves additional challenges, owing to focusing problems and motion blurring. Multiple frames under varying spatial or temporal settings can acquire additional information, which can be used to achieve improved classification performance. This study investigates the effectiveness of multi-frame decision-level fusion with photon-counting linear discriminant analysis. Multiple frames generate multiple scores for each class. The fusion process comprises three stages: score normalization, score validation, and score combination. Candidate scores are selected during the score validation process, after the scores are normalized. The score validation process removes bad scores that can degrade the final output. The selected candidate scores are combined using one of the following fusion rules: maximum, averaging, and majority voting. Degraded facial images are employed to demonstrate the robustness of multi-frame decision-level fusion in harsh environments. Out-of-focus and motion blurring point-spread functions are applied to the test images, to simulate long-distance acquisition. Experimental results with three facial data sets indicate the efficiency of the proposed decision-level fusion scheme.

• 보건행정학회지
• /
• 제13권1호
• /
• pp.98-115
• /
• 2003

Concerns about growing health insurance expenditures became a national Issue in 2001 when the National Health Insurance went into a deficit. Increases in spending for ambulatory care shared the largest portion of the problem. Methods and systems to control the spending should be developed and a system to measure case mix of providers is one of core components of the control system. The objectives of this article is to examine the feasibility of applying Korean Diagnosis Related Groups (KDRGs) to classify health insurance claims for ambulatory care and to identify problem areas of the classification. A database of 11,586,270 claims for ambulatory care delivered during January 2002 was obtained for the study, and the final number of claims analyzed was 8,319,494 after KDRG numbers were assigned to the data and records with an error KDRG were excluded from the study. The unit of analysis was a claim and resource use was measured by the sum of charges incurred during a month at a department of a hospital of at a clinic. Within group variance was assessed by th coefficient of variation (CV), and the classification accuracy was evaluated by the variance reduction achieved by the KDRG classification. The analyses were performed on both all and non-outlier data, and on a subset of the database to examine the validity of study results. Data were assigned to 787 KDRGs among 1,244 KDRGs defined in the classification system. For non-outlier data, 77.4％ of KDRGs had a CV of charges from tertiary care hospitals less than 100％ and 95.43％ of KDRGs for data from clinics. The variance reduction achieved by the KDRG classification was 40.80％ for non-outlier claims from tertiary care hospitals, 51.98％ for general hospitals, 40.89% for hospitals, and 54.99% for clinics. Similar results were obtained from the analyses performed on a subset of the study database. The study results indicated that KDRGs developed for a classification of inpatient care could be used for ambulatory care, although there were areas where the classification should be refined. Its power to predict tile resource utilization showed a potential for its application to measure case mix of providers for monitoring and managing delivery of ambulatory care. The issue concerning the quality of diagnostic information contained in insurance claims remains to be improved, and significance of future studies for other classification systems based on visits or episodes is guaranteed.

• 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.

• 정보처리학회논문지:컴퓨터 및 통신 시스템
• /
• 제7권1호
• /
• pp.1-8
• /
• 2018

동영상 데이터는 시간에 따른 정보는 물론이고, 많은 정보량과 함께 잡음도 포함하고 있기 때문에 이에 대한 간단한 표현을 학습하는 것은 쉽지 않다. 본 연구에서는 이와 같은 동영상 데이터를 추상적이면서 보다 간단하게 표현할 수 있는 순차 데이터간의 유사도 표현 방법과 딥러닝 학습방법을 제안한다. 이는 동영상을 구성하는 이미지 데이터 벡터들 사이의 유사도를 내적으로 표현할 때 그것들이 서로 최대한의 정보를 가질 수 있도록 하는 함수를 구하고 학습하는 것이다. 실제 데이터를 통하여 제안된 방법이 기존의 동영상 분류 방법들보다도 뛰어난 분류 성능을 보임을 확인하였다.

• Journal of Information Processing Systems
• /
• 제20권1호
• /
• pp.13-23
• /
• 2024

Music brings pleasure and relaxation to people. Therefore, it is necessary to classify musical genres based on scenes. Identifying favorite musical genres from massive music data is a time-consuming and laborious task. Recent studies have suggested that machine learning algorithms are effective in distinguishing between various musical genres. However, meeting the actual requirements in terms of accuracy or timeliness is challenging. In this study, a hybrid machine learning model that combines a deep residual auto-encoder (DRAE) and support vector machine (SVM) for musical genre recognition was proposed. Eight manually extracted features from the Mel-frequency cepstral coefficients (MFCC) were employed in the preprocessing stage as the hybrid music data source. During the training stage, DRAE was employed to extract feature maps, which were then used as input for the SVM classifier. The experimental results indicated that this method achieved a 91.54% F1-score and 91.58% top-1 accuracy, outperforming existing approaches. This novel approach leverages deep architecture and conventional machine learning algorithms and provides a new horizon for musical genre classification tasks.

• 대한임베디드공학회논문지
• /
• 제12권5호
• /
• pp.277-286
• /
• 2017

This paper introduces mathematical modelling and control algorithm of expendable mine disposal vehicle. This vehicle has two longitudinal thrusters, one vertical thruster and internal mass moving system which can control pitch rate. Also, the vehicle has an optical camera and forward looking sonar for underwater mine detection and classification. The vehicle is controlled via an optical cable connected with operating console on the mother ship. We describe the vehicle's 6DOF dynamic model and controller which can track the desired trajectory for the way-point tracking. These simulation results shows guidance and maneuvering performance which has other sensor data or not.

• 산업경영시스템학회지
• /
• 제41권2호
• /
• pp.56-64
• /
• 2018

In recent years, imbalanced data is one of the most important and frequent issue for quality control in industrial field. As an example, defect rate has been drastically reduced thanks to highly developed technology and quality management, so that only few defective data can be obtained from production process. Therefore, quality classification should be performed under the condition that one class (defective dataset) is even smaller than the other class (good dataset). However, traditional multi-class classification methods are not appropriate to deal with such an imbalanced dataset, since they classify data from the difference between one class and the others that can hardly be found in imbalanced datasets. Thus, one-class classification that thoroughly learns patterns of target class is more suitable for imbalanced dataset since it only focuses on data in a target class. So far, several one-class classification methods such as one-class support vector machine, neural network and decision tree there have been suggested. One-class support vector machine and neural network can guarantee good classification rate, and decision tree can provide a set of rules that can be clearly interpreted. However, the classifiers obtained from the former two methods consist of complex mathematical functions and cannot be easily understood by users. In case of decision tree, the criterion for rule generation is ambiguous. Therefore, as an alternative, a new one-class classifier using hyper-rectangles was proposed, which performs precise classification compared to other methods and generates rules clearly understood by users as well. In this paper, we suggest an approach for improving the limitations of those previous one-class classification algorithms. Specifically, the suggested approach produces more improved one-class classifier using hyper-rectangles generated by using Gaussian function. The performance of the suggested algorithm is verified by a numerical experiment, which uses several datasets in UCI machine learning repository.

• 한국정보시스템학회지:정보시스템연구
• /
• 제28권3호
• /
• pp.249-276
• /
• 2019

Purpose The purpose of this study is to develop a prediction model and decision rules for the elderly's suicidal ideation based on the Korean Welfare Panel survey data. By utilizing this data, we obtained many decision rules to predict the elderly's suicide ideation. Design/methodology/approach This study used classification analysis to derive decision rules to predict on the basis of decision tree technique. Weka 3.8 is used as the data mining tool in this study. The decision tree algorithm uses J48, also known as C4.5. In addition, 66.6% of the total data was divided into learning data and verification data. We considered all possible variables based on previous studies in predicting suicidal ideation of the elderly. Finally, 99 variables including the target variable were used. Classification analysis was performed by introducing sampling technique through backward elimination and data balancing. Findings As a result, there were significant differences between the data sets. The selected data sets have different, various decision tree and several rules. Based on the decision tree method, we derived the rules for suicide prevention. The decision tree derives not only the rules for the suicidal ideation of the depressed group, but also the rules for the suicidal ideation of the non-depressed group. In addition, in developing the predictive model, the problem of over-fitting due to the data imbalance phenomenon was directly identified through the application of data balancing. We could conclude that it is necessary to balance the data on the target variables in order to perform the correct classification analysis without over-fitting. In addition, although data balancing is applied, it is shown that performance is not inferior in prediction rate when compared with a biased prediction model.

• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2022년도 춘계학술대회
• /
• pp.469-472
• /
• 2022

최근 지능형 교통 시스템의 발전에 따라 딥러닝을 기술을 적용한 다양한 기술들이 활용되고 있다. 도로를 주행하는 불법 차량 및 범죄 차량 단속을 위해서는 차량 종류를 정확히 판별할 수 있는 차종 분류 시스템이 필요하다. 본 연구는 YOLO(You Only Look Once)를 이용하여 이동식 차량 단속 시스템에 최적화된 차종 분류 시스템을 제안한다. 제안 시스템은 차량을 승용차, 경·소·중형 승합차, 대형 승합차, 화물차, 이륜차, 특수차, 건설기계, 7가지 클래스로 구분하여 탐지하기 위해 단일 단계 방식의 객체 탐지 알고리즘 YOLOv5를 사용한다. 인공지능 기술개발을 위하여 한국과학기술연구원에서 구축한 약 5천 장의 국내 차량 이미지 데이터를 학습 데이터로 사용하였다. 한 대의 카메라로 정면과 측면 각도를 모두 인식할 수 있는 차종 분류 알고리즘을 적용한 지정차로제 단속 시스템을 제안하고자 한다.

• International Journal of Control, Automation, and Systems
• /
• 제4권2호
• /
• pp.187-196
• /
• 2006

A new GA-based methodology using information granules is suggested for the construction of fuzzy classifiers. The proposed scheme consists of three steps: selection of information granules, construction of the associated fuzzy sets, and tuning of the fuzzy rules. First, the genetic algorithm (GA) is applied to the development of the adequate information granules. The fuzzy sets are then constructed from the analysis of the developed information granules. An interpretable fuzzy classifier is designed by using the constructed fuzzy sets. Finally, the GA is utilized for tuning of the fuzzy rules, which can enhance the classification performance on the misclassified data (e.g., data with the strange pattern or on the boundaries of the classes). To show the effectiveness of the proposed method, an example, the classification of the Iris data, is provided.