• Journal of Korea Multimedia Society
• /
• v.23 no.12
• /
• pp.1486-1495
• /
• 2020

The analysis of digital microscopy images plays a vital role in computer-aided diagnosis (CAD) and prognosis. The main purpose of this paper is to develop a machine learning technique to predict the histological grades in prostate biopsy. To perform a multiclass classification, an AI-based deep learning algorithm, a multichannel convolutional neural network (MCCNN) was developed by connecting layers with artificial neurons inspired by the human brain system. The histological grades that were used for the analysis are benign, grade 3, grade 4, and grade 5. The proposed approach aims to classify multiple patterns of images extracted from the whole slide image (WSI) of a prostate biopsy based on the Gleason grading system. The Multichannel Convolution Neural Network (MCCNN) model takes three input channels (Red, Green, and Blue) to extract the computational features from each channel and concatenate them for multiclass classification. Stain normalization was carried out for each histological grade to standardize the intensity and contrast level in the image. The proposed model has been trained, validated, and tested with the histopathological images and has achieved an average accuracy of 96.4%, 94.6%, and 95.1%, respectively.

• Quality Improvement in Health Care
• /
• v.12 no.2
• /
• pp.84-92
• /
• 2006

Background : To examine the difference between ICD-10 and The Korean standard classification of disease(oriental medicine), and to aim at improve the practical use as statistical data. It is one of the reason of disease classification. On that account we convert the many to many correspondence presenting classification of oriental medicine into many to one correspondence. Method : The study tracked out 155 patients discharged from the university hospital which is located in Gyeonggi Province and managing hospital and oriental medicine hospital from July to October this year. The period of this study was from August 1 to November 18. We compared correspondence between the two services' diagnosis(hospital services and oriental medicine hospital services) at the same time and attempted many to one correspondence classification. That is for production of statistical data. Result : We investigated the group which have had medical treatment experience of two kinds of services at the same time. The result of this investigation was that the same oriental medicine diagnosis used differently in western medicine diagnosis. 44.5% was accorded with western medicine diagnosis. Correspondence of the western medicine diagnose with the top of the Korean standard classification of disease(oriental medicine) list's western medicine diagnosis was 13.5%. For many to one correspondence classification for statistics, one western medicine diagnosis was selected for one oriental medicine diagnosis. In case of the main diagnosis(I sign) was not enough to explain oriental medicine diagnosis' characteristic, we chose multiple other diagnosis, so other diagnosis(II sign) about patient's cause of disease could be selected for supplement after we examined the patient's records. The statistics was possible with this many to one correspondence. Conclusion : The result of this study about correspondence between western medicine diagnoses and those of oriental medicine confirms that The Korean standard classification of disease(oriental medicine) is hard to be standardized with western medicine diagnosis. Therefore, according to this study, we use new many to one correspondence classification, multiple oriental medicine diagnoses with one ICD-10, which can be used by statistical data.

• Convergence Security Journal
• /
• v.23 no.5
• /
• pp.135-142
• /
• 2023

In addressing the prominent issues of climate change, resource scarcity, and environmental pollution associated with household waste, extensive research has been conducted on intelligent waste classification methods. These efforts range from traditional classification algorithms to machine learning and neural networks. However, challenges persist in effectively classifying waste in diverse environments and conditions due to insufficient datasets, increased complexity in neural network architectures, and performance limitations for real-world applications. Therefore, this paper proposes i-YOLOX as a solution for rapid classification and improved accuracy. The proposed model is evaluated based on network parameters, detection speed, and accuracy. To achieve this, a dataset comprising 10,000 samples of household waste, spanning 17 waste categories, is created. The i-YOLOX architecture is constructed by introducing the Involution channel convolution operator and the Convolution Branch Attention Module (CBAM) into the YOLOX structure. A comparative analysis is conducted with the performance of the existing YOLO architecture. Experimental results demonstrate that i-YOLOX enhances the detection speed and accuracy of waste objects in complex scenes compared to conventional neural networks. This confirms the effectiveness of the proposed i-YOLOX architecture in the detection and classification of multiple household waste objects.

• Water for future
• /
• v.10 no.1
• /
• pp.95-100
• /
• 1977

The empirical area-reduction method by W.M. Borland and C.R. Miller and its revised procedure by W.T. Moody were made of fitting the area and storage curves to the Van't Hul distributions. It should be noted that the reservoir is classified into one of the four standard types on the basis of the topographical feature of the reservoir in application of the method. In other words, this method did not take into account several considerafble factors affecting the mode of sediment deposition, but only the shape of the reservoir as a governign factor. This is why the method occasionally creates ambiguity in classification and accordingly leads to unexpected mode of deposition. This paper describes a generating an formula to decide the standard classification of four types Van's Hul distributions, taking into consideration quantitatively sediment-loss percent and capacity-inflow ratio as well as the shape of the reservoirs by multiple regression analysis using the least square method to get a better fit to the design curves. The result is expressed as $Y=-1.95+55.8X_1+0.14X_2+0.12X_3$ in which the the values of Y locate the standard type I through type IV in the range from ten to forty with the interval of ten. The regression analysis was correlated well with the standard errors of estimate of around two except for the case of the type IV. This formula does not give big difference from the Borland's work in general sityation, but it demonstrates acceptable results, giving somewhat precise replys for the specific reservoirs. Its application to the Soyang Lake, one of the largest reservoirs in the country, defined clearly the type II, while the original method located it in the boundary of the type II and type III.

• The Journal of Korea Robotics Society
• /
• v.19 no.3
• /
• pp.254-265
• /
• 2024

Pattern recognition for surface electromyogram (sEMG) suffers from its nonstationary and stochastic property. Although it can be relieved by acquiring new training data, it is not only time-consuming and burdensome process but also hard to set the standard when the data acquisition should be held. Therefore, we propose an adaptive sEMG pattern recognition algorithm using principal component analysis. The proposed algorithm finds the relationship between sEMG channels and extracts the optimal principal component. Based on the relative distance, the proposed algorithm determines whether to update the existing patterns or to register the new pattern. From the experimental result, it is shown that multiple patterns are generated from the sEMG data stream and they are highly related to the motion. Furthermore, the proposed algorithm has shown higher classification accuracy than k-nearest neighbor (k-NN) and support vector machine (SVM). We expect that the proposed algorithm is utilized for adaptive and long-lasting pattern recognition.

• Journal of the Korean Society for Library and Information Science
• /
• v.58 no.2
• /
• pp.269-288
• /
• 2024

To structure the KOS registry, it is necessary to select a classification system that suits the characteristics of the collected KOS. This study aimed to classify domestic KOS collected through various classification schems, and based on these results, provide insights for selecting a classification system when structuring the KOS registry. A total of 313 KOS data collected via web searches were categorized using five types of classification systems and a thesaurus, and the results were analyzed. The analysis indicated that for international linkage of the KOS registry, foreign classification systems should be applied, and for optimization with domestic knowledge resources or to cater to domestic researchers, domestic classification systems need to be applied. Additionally, depending on the field-specific characteristics of the KOS, research area KOS should apply classification systems based on academic disciplines, while public sector KOS should consider classification systems based on government functions. Lastly, it is necessary to strengthen the linkage between domestic and international KOS, which also requires the application of multiple classification systems.

• 한국정보디스플레이학회:학술대회논문집
• /
• 2002.08a
• /
• pp.856-859
• /
• 2002

A system for high-resolution analysis of defective elementary cell (R, G or B) on Flat Panel Display (FPD) is described. Based on multiple acquisitions of low-resolution shifted images of the display, our system doesn't require a high-resolution sensor neither tedious alignment of the display, and will remain up to date even facing an important increase of the display dimensions. Our process, highly automated and thus flexible and robust, is expected to perform a full analysis in less than 60s. It is mainly intended for production tests and display classification by manufacturers.

• Journal of the Korean Society of Costume
• /
• v.52 no.6
• /
• pp.51-69
• /
• 2002

The purpose of this study was to classify the body shapes. exclusive of size and corpulence factors of more than 60-year old elderly women by distinctions, and to investigate the visual effects of combination of ensemble suit details. For the body shape classification, the factor analysis and cluster analysis were performed : the mean value difference of numeral values for classified types were tested by ANOVA : and the follow-up test was conducted by the Duncan's multiple ranged test. The data analysis for visual effects evaluated by a multiple ranking test was analysed by mean. paired t-test, ANOVA and Duncan's multiple ranged test. The results are summarized as follows : 1. The followings are the types of body shape according to the shape factors of the front line of body for elderly women. The distinctions of the front li e of elderly women's body could be presumed; that was, Body typeⅠ was a comparatively well-balanced body type, Body type Ⅱ was close to an average body type. and Body type In was a severely corpulent body type. 2. The followings are the results on the physical visual effects inducing the constituents of clothing type. 1) The neckline·collar types of a jacket have a great influence on the visual effects of the upper body, and orderly. the tailored collar. soutien collar, and round neckline had positive influence on the visual effects in the upper body. 2) The pleat types of one-piece dress had positive influence on the visual effects in the lower body in the order of gored type, pleats type, and gathered type. Also. the balance in the lower body had more influence on the overall balance of the clothing compared to the constituents of clothing type such as neckline collar type or opening line. 3) It showed that whether there is the front opening line of a jacket influenced on the visual effects of all categories.

• Journal of Families and Better Life
• /
• v.10 no.1 s.19
• /
• pp.27-42
• /
• 1992

This study focused on consumer information search activity and consumer's post-purchase satisfaction. For these purpose, a survey was conducted suing questionaires on 430 homemakers that lived in seoul. Statistics used for data were Frequency Distribution. Percentile, Mean, One-way AAANOVA., Scheffe-test, T-test, Pearson's correlation. Multiple Regression Analysis and Multiple Classification Analysis. The major findings were ; 1) The level of each amount information search was lower than average. And the level of consumer's post-purchase satisfaction was a little higher than average. 2) On amount of "noncommercial-personal" information search, the influencing variables were desire to seek information, education, brand royalty in turn. These three variables explained 7% of dependent variable's variance. 3) On amount of "noncommercial-media" information search, the influencing variables were desire to seek information, amount of internal information, education, occupational status in turn. These variables explained 14% of dependent variable's variance. 4) On amount of "commercial-personal" information search, the influencing variable was desire to seek information, and this variable explained 3.1% of dependent variable'a variance. 5) On amount of "commercial-media" information search, the influencing variables were desire to seek information, education, amount of internal information in turn. These three variables explained 12.1% dependent variable's variance. 6) Resulting from multiple classification analysis, influencing variables on consumer's post-purchase satisfaction were amount of noncommercial-media information search and printed media search, and brand royalty. These three variables explained 9% of dependent variable's variance. Furthermore, througout all the subareas of consumer's satisfaction, the amount of noncommercial-media information search was the most influencing variable.

• Journal of Biomedical Engineering Research
• /
• v.31 no.6
• /
• pp.456-463
• /
• 2010

In the present study, we proposed a new subspace scanning algorithm to enhance the spatial resolution of electroencephalography (EEG) and magnetoencephalography(MEG) source localization. Subspace scanning algorithms, represented by the multiple signal classification (MUSIC) algorithm and the first principal vector (FINE) algorithm, have been widely used to localize asynchronous multiple dipolar sources in human cerebral cortex. The conventional MUSIC algorithm used principal component analysis (PCA) to extract the noise vector subspace, thereby having difficulty in discriminating two or more closely-spaced cortical sources. The FINE algorithm addressed the problem by using only a part of the noise vector subspace, but there was no golden rule to determine the number of noise vectors. In the present work, we estimated a non-orthogonal signal vector set using independent component analysis (ICA) instead of using PCA and performed the source scanning process in the signal vector subspace, not in the noise vector subspace. Realistic 2D and 3D computer simulations, which compared the spatial resolutions of various algorithms under different noise levels, showed that the proposed ICA-MUSIC algorithm has the highest spatial resolution, suggesting that it can be a useful tool for practical EEG/MEG source localization.