• Transactions of Materials Processing
• /
• v.31 no.4
• /
• pp.229-239
• /
• 2022

Convolution Neural Network(CNN) is a class of deep learning algorithms and can be used for image analysis. In particular, it has excellent performance in finding the pattern of images. Therefore, CNN is commonly applied for recognizing, learning and classifying images. In this study, the surface defect classification performance of Al 6061 extruded material using CNN-based algorithms were compared and evaluated. First, the data collection criteria were suggested and a total of 2,024 datasets were prepared. And they were randomly classified into 1,417 learning data and 607 evaluation data. After that, the size and quality of the training data set were improved using data augmentation techniques to increase the performance of deep learning. The CNN-based algorithms used in this study were VGGNet-16, VGGNet-19, ResNet-50 and DenseNet-121. The evaluation of the defect classification performance was made by comparing the accuracy, loss, and learning speed using verification data. The DenseNet-121 algorithm showed better performance than other algorithms with an accuracy of 99.13% and a loss value of 0.037. This was due to the structural characteristics of the DenseNet model, and the information loss was reduced by acquiring information from all previous layers for image identification in this algorithm. Based on the above results, the possibility of machine vision application of CNN-based model for the surface defect classification of Al extruded materials was also discussed.

• Journal of Korean Society for Quality Management
• /
• v.43 no.3
• /
• pp.273-288
• /
• 2015

Purpose: In this study, we apply environmental stress screening (ESS) to battle management system (BMS) of a tank and use the ESS profile based on production process data, guided by MIL-HDBK-781/344/2164. Methods: To optimize ESS Profile of the BMS of a tank, we estimate ESS model parameters (e.g., defect density, screening strength) using primary production failure reporting and corrective action system (FRACAS) data of military supply contract firm. Results: First, we collect the Primary production FRACAS data of military supply contract firm. Second, we compute curve fitting approach to find patent defect density and latent defect density using FRACAS data. Third, we solve the equation of Defect Density(patent defect density + latent defect density)($D_{IN}$) and Screening Strength(SS) Using second step data. As a result of analysis according to the order, we calculate $D_{IN}$(Temperature stress case : 74.02, Vibration stress : 10.252) and : SS(Temperature stress case : 0.4632, Vibration stress : 0.4142) and confirm the Condition II-D based on MIL-HDBK-344. According to Condition II-D, it is necessary to modify existing ESS profile through decreasing the $D_{IN}$ and increasing the SS. Conclusion: Identification of defect causes through ESS approach reduce defect densities for production. It provides feedback to a lessons-learned data base to avoid similar problems on next generation tank BMS.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.124-125
• /
• 2023

Defect detection during liquid crystal display (LCD) manufacturing has always been a critical challenge. This study aims to address this issue by proposing a data augmentation method based on generative adversarial networks (GAN) to improve defect identification accuracy in LCD production. By leveraging synthetically generated image data from GAN, we effectively augment the original dataset to make it more representative and diverse. This data augmentation strategy enhances the model's generalization capability and robustness on real-world data. Compared to traditional data augmentation techniques, the synthetic data from GAN are more realistic, diverse and broadly distributed. Experimental results demonstrate that training models with GAN-generated data combined with the original dataset significantly improves the detection accuracy of critical defects in LCD manufacturing, compared to using the original dataset alone. This study provides an effective data augmentation approach for intelligent quality control in LCD production.

• Journal of the Korean Institute of Electrical and Electronic Material Engineers
• /
• v.28 no.5
• /
• pp.320-325
• /
• 2015

This paper dealt with a defect identification algorithm which is based on single partial discharge (PD) pulse analysis in gas insulated structure. Four types of electrode systems such as a needle-plane, a plane-needle, a free particle and a crack inside spacer were fabricated to simulate defects in gas insulated switchgear (GIS). We measured single PD pulse by an oscilloscope with a sampling rate of 5 GS/s and a frequency bandwidth of 1 GHz. Data aquisition and signal processing were controlled by a LabVIEW program. Physical shapes of PD pulses were compared with kurtosis, skewness and time-based parameters as rising time, falling time and pulse-width. These parameters were analysed by an algorithm with a back propagation algorithm (BPA). By applying the algorithm, the identification rate was 97% for the needle-plane electrode, 96% for the plane-needle electrode, 91% for the free particle and 93% for the crack inside spacer. The results verified that the algorithm could identify the type of defects in GIS.

• Journal of Institute of Control, Robotics and Systems
• /
• v.18 no.9
• /
• pp.801-805
• /
• 2012

This paper proposes dynamic thresholding scheme for fingerprint identification. As a user authentication method by fingerprint recognition technology, verification method based on 1:1 matching was mainly used in the past, but identification method based on 1:N matching is generally used recently. The control of the value of FAR is very important in the application areas such as access control and time attendance systems. This paper proposes dynamic thresholding scheme which could properly control the value of FAR according to the field of applications and size of the fingerprints database.

• Journal of the Korean Society for Nondestructive Testing
• /
• v.34 no.1
• /
• pp.38-43
• /
• 2014

An experimental method based on flexural wave propagation is proposed for identification of structural damage in rail fastening systems. The vibration of a rail clamped and supported by viscoelastic pads is significantly influenced by dynamic support properties. Formation of a defect in the rail fastening system induces changes in the flexural wave propagation characteristics owning to the discontinuity in the structural properties. In this study, frequency-dependent support stiffness was measured to monitor this change by a transfer function method. The sensitivity of wave propagation on the defect was measured from the potential energy stored in a continuously supported rail. Further, the damage index was defined as a correlation coefficient between the change in the support stiffness and the sensitivity. The defect location was identified from the calculated damage index.

• Journal of Periodontal and Implant Science
• /
• v.46 no.4
• /
• pp.244-253
• /
• 2016

Purpose: The aim of this study was to characterize the healing in the grafted calvarial defects of rats after adjunctive hyperbaric oxygen therapy. Methods: Twenty-eight male Sprague-Dawley rats (body weight, 250-300 g) were randomly divided into two treatment groups: with hyperbaric oxygen therapy (HBO; n=14) and without HBO (NHBO; n=14). Each group was further subdivided according to the bone substitute applied: biphasic calcium phosphate (BCP; n=7) and surface-modified BCP (mBCP; n=7). The mBCP comprised BCP coated with Escherichia-coli-derived recombinant human bone morphogenetic protein-2 (ErhBMP-2) and epigallocatechin-3-gallate (EGCG). Two symmetrical circular defects (6-mm diameter) were created in the right and left parietal bones of each animal. One defect was assigned as a control defect and received no bone substitute, while the other defect was filled with either BCP or mBCP. The animals were allowed to heal for 4 weeks, during which those in the HBO group underwent 5 sessions of HBO. At 4 weeks, the animals were sacrificed, and the defects were harvested for histologic and histomorphometric analysis. Results: Well-maintained space was found in the grafted groups. Woven bone connected to and away from the defect margin was formed. More angiogenesis was found with HBO and EGCG/BMP-2 (P<0.05). None of the defects achieved complete defect closure. Increased new bone formation with HBO or EGCG/BMP-2 was evident in histologic evaluation, but it did not reach statistical significance in histometric analysis. A synergic effect between HBO and EGCG/BMP-2 was not found. Conclusions: Within the limitations of this study, the present findings indicate that adjunctive HBO and EGCG/BMP-2 could be beneficial for new bone formation in rat calvarial defects.

• Current Research on Agriculture and Life Sciences
• /
• v.30 no.1
• /
• pp.1-11
• /
• 2012

We identified cdf based on screening of the Arabidopsis cDNA library for functional suppressors of the AtBI-1 (a gene described to suppress the cell death induced by Bax gene expression in yeast). The cdf was located on Chr. V and was composed of 5 exons and 4 introns. It encodes a protein of 258 amino acid residues with a molecular weight of 28.8 kDa. The protein has 3 transmembrane domains in the C-terminal region. The cdf has one homologue, named cdf2, which was found in Arabidopsis. Like cdf, cdf2 also induced growth defect in yeast. The effect of the cell growth defect factor was somewhat lower than Bax. cdf could arrest the growth of yeast. Its localization to the nucleus was essential for the suppression of yeast cell proliferation. Morphological abnormality of intracellular network, which is a hallmark of AtBI-1, was attenuated by expression of cdf.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.2 no.1
• /
• pp.89-93
• /
• 2002

This paper introduces a vision-based on-line fabric inspection methodology of woven textile fabrics. Current procedure for determination of fabric defects in the textile industry is performed by human in the off-line stage. The advantage of the on-line inspection system is not only defect detection and identification, but also 벼ality improvement by a feedback control loop to adjust set-points. The proposed inspection system consists of hardware and software components. The hardware components consist of CCD array cameras, a frame grabber and appropriate illumination. The software routines capitalize upon vertical and horizontal scanning algorithms characteristic of a particular deflect. The signal to noise ratio (SNR) calculation based on the results of the wavelet transform is performed to measure any deflects. The defect declaration is carried out employing SNR and scanning methods. Test results from different types of defect and different style of fabric demonstrate the effectiveness of the proposed inspection system.

• Transactions of the Korean Society of Machine Tool Engineers
• /
• v.13 no.5
• /
• pp.87-95
• /
• 2004

From the results of AE(Acoustic Emission) source location occurred by the spot exciting as suggested in this research, it has been confirmed that AE technique is quite fruitful in figuring out the location of the occurrence, form, size and direction of the defects. As the results of examining the distribution of event for the angle of crack $\alpha$ to Xs and Ys, as the increases from $0^{\circ}$ ~ $90^{\circ}$, gradually changes its width from the axis Xs to the axis Ys. So event appears approximately similar in its size at the angle of crack $\alpha$=$45^{\circ}$, yet opposite when $\alpha$ is lager. It is believed that this is a phenomenon where its crack legnth $\alpha$, assumed as a planar defect, is to be prcjected toward the direction with a larger size. Thus, it is expected that the application of the experimental method suggested in this study would make it possible to identify the location of the defect in the material in the nondestructive way.