• Journal of Industrial Convergence
• /
• v.21 no.7
• /
• pp.19-27
• /
• 2023

As the secondary battery market expands, the process of producing laterite ore using the rotary kiln and electric furnace method is expanding worldwide. As ESG management expands, the management of air pollutants such as nitrogen oxides in exhaust gases is strengthened. The rotary kiln, one of the main facilities of the pyrometallurgy process, is a facility for drying and preliminary reduction of ore, and it generate nitrogen oxides, thus prediction of nitrogen oxide is important. In this study, LSTM for regression prediction and LightGBM for classification prediction were used to predict and then model optimization was performed using AutoML. When applying LSTM, the predicted value after 5 minutes was 0.86, MAE 5.13ppm, and after 40 minutes, the predicted value was 0.38 and MAE 10.84ppm. As a result of applying LightGBM for classification prediction, the test accuracy rose from 0.75 after 5 minutes to 0.61 after 40 minutes, to a level that can be used for actual operation, and as a result of model optimization through AutoML, the accuracy of the prediction after 5 minutes improved from 0.75 to 0.80 and from 0.61 to 0.70. Through this study, nitrogen oxide prediction values can be applied to actual operations to contribute to compliance with air pollutant emission regulations and ESG management.

• Journal of Digital Convergence
• /
• v.10 no.11
• /
• pp.383-387
• /
• 2012

Development of automated document management system specified for scanned invoice images suffers from rigorous accuracy requirements for extraction of monetary data, which necessiate automatic validation on the extracted values for a generative invoice table model. Use of certain internal constraints such as "amount = unit price times quantity" is typical implementation. In this paper, we propose a noble invoice information extraction model with improved auto-validation method by utilizing table header detection and column classification.

• Journal of Information Processing Systems
• /
• v.15 no.1
• /
• pp.100-115
• /
• 2019

Malicious code distribution on the Internet is one of the most critical Internet-based threats and distribution technology has evolved to bypass detection systems. As a new defense against the detection bypass technology of malicious attackers, this study proposes the automated tracing of malicious websites in a malware distribution network (MDN). The proposed technology extracts automated links and classifies websites into malicious and normal websites based on link structure. Even if attackers use a new distribution technology, website classification is possible as long as the connections are established through automated links. The use of a real web-browser and proxy server enables an adequate response to attackers' perception of analysis environments and evasion technology and prevents analysis environments from being infected by malicious code. The validity and accuracy of the proposed method for classification are verified using 20,000 links, 10,000 each from normal and malicious websites.

• Journal of Convergence for Information Technology
• /
• v.11 no.6
• /
• pp.49-54
• /
• 2021

With the recent development of smart healthcare technology, interest in daily diseases is increasing. However, healthcare data has an imbalance between positive and negative data. This is caused by the difficulty of collecting data because there are relatively many people who are not patients compared to patients with certain diseases. Data imbalances need to be adjusted because they affect performance in ongoing learning during disease prediction and analysis. Therefore, in this paper, We replace missing values through multiple imputation in detection models to determine whether they are prevalent or not, and resolve data imbalances through over-sampling. Based on AutoML using preprocessed data, We generate several models and select top 3 models to generate ensemble models.

• Korean Journal of Radiology
• /
• v.22 no.11
• /
• pp.1764-1776
• /
• 2021

Objective: This study aimed to validate a deep learning-based fully automatic calcium scoring (coronary artery calcium [CAC]_auto) system using previously published cardiac computed tomography (CT) cohort data with the manually segmented coronary calcium scoring (CAC_hand) system as the reference standard. Materials and Methods: We developed the CAC_auto system using 100 co-registered, non-enhanced and contrast-enhanced CT scans. For the validation of the CAC_auto system, three previously published CT cohorts (n = 2985) were chosen to represent different clinical scenarios (i.e., 2647 asymptomatic, 220 symptomatic, 118 valve disease) and four CT models. The performance of the CAC_auto system in detecting coronary calcium was determined. The reliability of the system in measuring the Agatston score as compared with CAC_hand was also evaluated per vessel and per patient using intraclass correlation coefficients (ICCs) and Bland-Altman analysis. The agreement between CAC_auto and CAC_hand based on the cardiovascular risk stratification categories (Agatston score: 0, 1-10, 11-100, 101-400, > 400) was evaluated. Results: In 2985 patients, 6218 coronary calcium lesions were identified using CAC_hand. The per-lesion sensitivity and false-positive rate of the CAC_auto system in detecting coronary calcium were 93.3% (5800 of 6218) and 0.11 false-positive lesions per patient, respectively. The CAC_auto system, in measuring the Agatston score, yielded ICCs of 0.99 for all the vessels (left main 0.91, left anterior descending 0.99, left circumflex 0.96, right coronary 0.99). The limits of agreement between CAC_auto and CAC_hand were 1.6 ± 52.2. The linearly weighted kappa value for the Agatston score categorization was 0.94. The main causes of false-positive results were image noise (29.1%, 97/333 lesions), aortic wall calcification (25.5%, 85/333 lesions), and pericardial calcification (24.3%, 81/333 lesions). Conclusion: The atlas-based CAC_auto empowered by deep learning provided accurate calcium score measurement as compared with manual method and risk category classification, which could potentially streamline CAC imaging workflows.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.36 no.2C
• /
• pp.78-85
• /
• 2011

In this paper we proposed an intelligent PTZ camera system which detects, classifies and tracks moving objects. If a moving object is detected, features are extracted for classification and then realtime tracking follows. We used GMM for detection followed by shadow removal. Legendre moment is used for classification. Without auto focusing, we can control the PTZ camera movement by using center points of the image and object's direction, distance and velocity. To implement the realtime system, we used TI DM6446 Davinci processor. Throughout the experiment, we obtained system's high performance in classification and tracking both at vehicle's normal and high speed motion.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.19 no.3
• /
• pp.294-303
• /
• 2008

ISAR(Inverse Synthetic Aperture Radar) images represent the 2-D(two-dimensional) spatial distribution of RCS (Radar Cross Section) of an object, and they can be applied to the problem of target identification. A traditional approach to ISAR imaging is to use a 2-D IFFT(Inverse Fast Fourier Transform). However, the 2-D IFFT results in low resolution ISAR images especially when the measured frequency bandwidth and angular region are limited. In order to improve the resolution capability of the Fourier transform, various high-resolution spectral estimation approaches have been applied to obtain ISAR images, such as AR(Auto Regressive), MUSIC(Multiple Signal Classification) or Modified MUSIC algorithms. In this study, these high-resolution spectral estimators as well as 2-D IFFT approach are combined with a recently developed ISAR image classification algorithm, and their performances are carefully analyzed and compared in the framework of radar target recognition.

• Journal of Korean Institute of Industrial Engineers
• /
• v.28 no.2
• /
• pp.147-161
• /
• 2002

In novelty detection, one attempts to discriminate abnormal patterns from normal ones. Novelty detection is quite difficult since, unlike usual two class classification problems, only normal patterns are available for training. Auto-Associative Multi-Layer Perceptron (AAMLP) has been shown to provide a good performance based upon the property that novel patterns usually have larger auto-associative errors. In this paper, we give a mathematical analysis of 2-layer AAMLP's output characteristics and empirical results of 2-layer and 4-layer AAMLPs. Various activation functions such as linear, saturated linear and sigmoid are compared. The 2-layer AAMLPs cannot identify non-linear boundaries while the 4-layer ones can. When the data distribution is multi-modal, then an ensemble of AAMLPs, each of which is trained with pre-clustered data is required. This paper contributes to understanding of AAMLP networks and leads to practical recommendations regarding its use.

• Industrial Engineering and Management Systems
• /
• v.13 no.4
• /
• pp.454-462
• /
• 2014

Retention of possible churning customer is one of the most important issues in customer relationship management, so companies try to predict churn customers using their large-scale high-dimensional data. This study focuses on dealing with large data sets by reducing the dimensionality. By using six different dimension reduction methods-Principal Component Analysis (PCA), factor analysis (FA), locally linear embedding (LLE), local tangent space alignment (LTSA), locally preserving projections (LPP), and deep auto-encoder-our experiments apply each dimension reduction method to the training data, build a classification model using the mapped data and then measure the performance using hit rate to compare the dimension reduction methods. In the result, PCA shows good performance despite its simplicity, and the deep auto-encoder gives the best overall performance. These results can be explained by the characteristics of the churn prediction data that is highly correlated and overlapped over the classes. We also proposed a simple out-of-sample extension method for the nonlinear dimension reduction methods, LLE and LTSA, utilizing the characteristic of the data.

• ETRI Journal
• /
• v.44 no.4
• /
• pp.613-623
• /
• 2022

Autism spectrum disorder (ASD) is a developmental disorder associated with cognitive and neurobehavioral disorders. It affects the person's behavior and performance. Autism affects verbal and non-verbal communication in social interactions. Early screening and diagnosis of ASD are essential and helpful for early educational planning and treatment, the provision of family support, and for providing appropriate medical support for the child on time. Thus, developing automated methods for diagnosing ASD is becoming an essential need. Herein, we investigate using various machine learning methods to build predictive models for diagnosing ASD in children using facial images. To achieve this, we used an autistic children dataset containing 2936 facial images of children with autism and typical children. In application, we used classical machine learning methods, such as support vector machine and random forest. In addition to using deep-learning methods, we used a state-of-the-art method, that is, automated machine learning (AutoML). We compared the results obtained from the existing techniques. Consequently, we obtained that AutoML achieved the highest performance of approximately 96% accuracy via the Hyperpot and tree-based pipeline optimization tool optimization. Furthermore, AutoML methods enabled us to easily find the best parameter settings without any human efforts for feature engineering.