• Journal of the Korean Society of Industry Convergence
• /
• v.23 no.2_2
• /
• pp.163-171
• /
• 2020

The condition of tire tread is a key parameter closely related to the driving safety of a vehicle, which affects the contact force of the tire for braking, accelerating and cornering. The major factor influencing the contact force is tread wear, and the more tire tread wears out, the higher risk of losing control of a vehicle exits. The tire tread condition is generally checked by visual inspection that can be easily forgotten. In this paper, we propose the intelligent tire (iTire) system that consists of an acceleration sensor, a wireless signal transmission unit and a tread classifier. In addition, we also presents classification algorithm that transforms the acceleration signal into the frequency domain and extracts the features of several frequency bands as inputs to an artificial neural network. The artificial neural network for classifying tire wear was designed with an Multiple Layer Perceptron (MLP) model. Experiments showed that tread wear classification accuracy was over 80%.

• Structural Engineering and Mechanics
• /
• v.25 no.1
• /
• pp.39-52
• /
• 2007

This paper describes an experimental study on structural health monitoring of a 1:3-scaled one-story concrete frame subjected to seismic damage and retrofit. The structure is tested on a shaking table by exerting successively enhanced earthquake excitations until severe damage, and then retrofitted using fiber-reinforced polymers (FRP). The modal properties of the tested structure at trifling, moderate, severe damage and strengthening stages are measured by subjecting it to a small-amplitude white-noise excitation after each earthquake attack. Making use of the measured global modal frequencies and a validated finite element model of the tested structure, a neural network method is developed to quantitatively identify the stiffness reduction due to damage and the stiffness enhancement due to strengthening. The identification results are compared with 'true' damage severities that are defined and determined based on visual inspection and local impact testing. It is shown that by the use of FRP retrofit, the stiffness of the severely damaged structure can be recovered to the level as in the trifling damage stage.

• Korean Journal of Clinical Pharmacy
• /
• v.25 no.3
• /
• pp.131-137
• /
• 2015

Background: The protective effect of metformin against breast cancer is inconclusive. Objective: To evaluate the effect of metformin on breast cancer risk and mortality in patients with type 2 diabetes. Method: A comprehensive literature search was performed for pertinent articles published prior to June 30, 2014, using PubMed and EMBASE. Study heterogeneity was estimated with $I^2$ statistic. The data from the included studies were pooled and weighted by random-effects model. The quality of each included study was assessed on the basis of the 9-star Newcastle-Ottawa Scale and publication bias was evaluated by visual inspection of a funnel plot. Results: Ten studies were included in the meta-analysis of the association of metformin and breast cancer risk. By synthesizing the data from the studies, the pooled odds ratio (OR) was 0.72 (95% CI: 0.59, 0.87) (p = 0.0005). Three cohort studies were included for meta-analysis of the association between metformin and breast cancer-related mortality. Metformin was associated with a significant decrease in mortality (Risk ratio: 0.68; 95% CI: 0.51, 0.90, p = 0.007). Conclusion: The present meta-analysis suggests that metformin appears to be associated with a lower risk of breast cancer incidence and mortality in patients with type 2 diabetes.

• Bulletin of the Korean Chemical Society
• /
• v.35 no.7
• /
• pp.2097-2108
• /
• 2014

Chemical feature based quantitative pharmacophore models were generated using the HypoGen module implemented in DS2.5. The best hypothesis, Hypo1, which was characterized by the highest correlation coefficient (0.96), the highest cost difference (61.60) and the lowest RMSD (0.74), consisted of one hydrogen bond acceptor, one hydrogen bond donor, one hydrophobic and one ring aromatic. The reliability of Hypo1 was validated on the basis of cost analysis, test set, Fischer's randomization method and GH test method. The validated Hypo1 was used as a 3D search query to identify novel inhibitors. The screened molecules were further refined by employing ADMET, docking studies and visual inspection. Three compounds with novel scaffolds were selected as the most promising candidates for the designing of Mcl-1 antagonists. Finally, a 10 ns molecular dynamics simulation was carried out on the complex of receptor and the retrieved ligand to demonstrate that the binding mode was stable during the MD simulation.

• The Bulletin of The Korean Astronomical Society
• /
• v.43 no.1
• /
• pp.54.3-54.3
• /
• 2018

In this study, we perform a statistical investigation on the kinematic classication of 4264 coronal mass ejections (CMEs) from 1996 to 2015 observed by SOHO/LASCO C3. Using the constant acceleration model, we classify these CMEs into three groups; deceleration, constant velocity, and acceleration motion. For this, we devise four dierent classication methods by acceleration, fractional speed variation, height contribution, and visual inspection. Our major results are as follows. First, the fractions of three groups depend on the method used. Second, about half of the events belong to the groups of acceleration and deceleration. Third, the fractions of three motion groups as a function of CME speed classied by the last three methods are consistent with one another. Fourth, according to the last three methods, the fraction of acceleration motion decreases as CME speed increases, while the fractions of other motions increase with speed. In addition, the acceleration motions are dominant in low speed CMEs whereas the constant velocity motions are dominant in high speed CMEs.

• Journal of Korea Multimedia Society
• /
• v.18 no.12
• /
• pp.1519-1527
• /
• 2015

TFT-LCD automatic defect inspection system for detecting defects in place of the visual tester does pre-processing, candidate defect pixel detection, and recognition and classification through a blob analysis. An over-detection result of defects acts as an undue burden of blob analysis for recognition and classification. In this paper, we propose defect detection method based on the histogram distribution modeling of TFT-LCD image to minimize over-detection of candidate defective pixels. Primary defect candidate pixels are detected estimating the skewness of the luminance distribution histogram of the background pixels. Based on the detected defect pixels, the defective pixels other than noise pixels are detected using the distribution histogram model of the local area. Experimental results confirm that the proposed method shows an excellent defect detection result on the image containing the various types of defects and the reduction of the degree of over-detection as well.

• The Bulletin of The Korean Astronomical Society
• /
• v.36 no.1
• /
• pp.86.2-86.2
• /
• 2011

The faint wing-like features at velocities beyond the velocity boundaries of the Galactic rotation (Forbidden-Velocity Wings, FVWs) in the large-scale position-velocity diagrams of the HI surveys are thought to be associated with dynamical Galactic events. The primary candidates of these FVWs are rapidly expanding HI shells of the old Galactic supernova remnants (SNRs), which are too faint to be visible in other frequencies. The unprecedented sensitivity and resolution of the I-GALFA survey enable detection of "all" HI shells of Galactic SNRs at forbidden velocities predicted by Koo and Kang (2004). Therefore, comparing the distribution of the FVWs visible in the I-GALFA survey and that of the model will improve our understanding on the interstellar medium and the evolution of SNRs. We have been developing an automatic searching algorithm for FVWs in the I-GALFA survey to minimize the selection effects of visual inspection used in the previous FVW study. We present the searching mechanism for FVWs and the statistical properties of the automatically searched FVWs. Also, we discuss the similarities and the differences between the distribution of the identified FVWs and that of the SNRs predicted by Koo and Kang (2004).

• The Journal of the Korean dental association
• /
• v.52 no.8
• /
• pp.456-463
• /
• 2014

Caries Management by Risk Assessment (CAMBRA), published by California Dental Association in 2003, is a customized caries care system that classifies individuals' caries risk into 4 risk groups based on objective evidences and provides chemical treatments targeted for each caries risk level. However, this system was not only developed but also optimized for situation in the United States, resulting into many limitations to be used in Korea, and thus Korean CAMBRA (K-CAMBRA) that considers the clinical situation in Korea needs to be developed. K-CAMBRA includes various techniques that are newly developed in order to overcome the limitations. First, Q-ray, a new optical technology, is utilized in order to avoid the subjectivity of visual inspection during assessment of disease indicators and risk factors. Moreover, Cariview? that reflects the paradigm shift in cariology as a new form of caries assessment kit is used. In addition, considering the situation in Korea, where it is impossible to use high concentration fluoride product, Oral pack with a customized tray is added to increase the contact time of chemical substance. CAMBRA is believed to be the key clinical tool that overcomes the limitations of the paradigm of the conventional restoration-based surgical model of dentistry. Furthermore, it can be expected that Korean dentists can act as oral physicians who are able to control and care individuals' caries risk rather than operative experts who only care about the outcome of caries.

• Journal of The Korean Astronomical Society
• /
• v.55 no.3
• /
• pp.67-74
• /
• 2022

In this study, we perform a statistical investigation of the kinematic classification of 4,264 coronal mass ejections (CMEs) from 1996 to 2015 observed by SOHO/LASCO C3. Using the constant acceleration model, we classify these CMEs into three groups: deceleration, constant velocity, and acceleration motion. For this, we devise three different classification methods using fractional speed variation, height contribution, and visual inspection. The main results of this study can be summarized as follows. First, the fractions of three groups depend on the method used. Second, about half of the events belong to the groups of acceleration and deceleration. Third, the fractions of three motion groups as a function of CME speed are consistent with one another. Fourth, the fraction of acceleration motion decreases as CME speed increases, while the fractions of other motions increase with speed. In addition, the acceleration motions are dominant in low speed CMEs whereas the constant velocity motions are dominant in high speed CMEs.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.17 no.3
• /
• pp.129-138
• /
• 2022

Collecting and labeling sufficient training data, which is essential to deep learning-based visual inspection, is difficult for manufacturers to perform because it is very expensive. This paper presents a steel plate surface defect detection system with industrial-grade detection performance by training a small amount of steel plate surface images consisting of labeled and non-labeled data. To overcome the problem of lack of training data, we propose two data augmentation techniques: program-based augmentation, which generates defect images in a geometric way, and generative model-based augmentation, which learns the distribution of labeled data. We also propose a 4-step semi-supervised learning using pseudo labels and consistency training with fixed-size augmentation in order to utilize unlabeled data for training. The proposed technique obtained about 99% defect detection performance for four defect types by using 100 real images including labeled and unlabeled data.