In this paper, a method for improving the defect classification performance in steel plate surface has been studied, based on DWT(discrete wavelet transform) and SVM(support vector machine). Surface images of the steel plate have low contrast, uneven, and featureless, so that the contrast between defect and defect-free regions is not discriminated. These characteristics make it difficult to extract the feature of the surface defect image. In order to improve the characteristics of these images, a synthetic images based on discrete wavelet transform are modeled. Using the synthetic images, edge-based features are extracted and also geometrical features are computed. SVM was configured in order to classify defect images using extracted features. As results of the experiment, the support vector machine based classifier showed good classification performance of 94.3%. The proposed classifier is expected to contribute to the key element of inspection process in smart factory.

Enormous methods have been proposed for the detection and segmentation of blur and non-blur regions of the images. Due to the limited available information about the blur type, scenario and the level of blurriness, detection and segmentation is a challenging task. Hence, the performance of the blur measure operators is an essential factor and needs improvement to attain perfection. In this paper, we propose an effective blur measure based on the local binary pattern (LBP) with the adaptive threshold for blur detection. The sharpness metric developed based on LBP uses a fixed threshold irrespective of the blur type and level which may not be suitable for images with large variations in imaging conditions and blur type and level. Contradictory, the proposed measure uses an adaptive threshold for each image based on the image and the blur properties to generate an improved sharpness metric. The adaptive threshold is computed based on the model learned through the support vector machine (SVM). The performance of the proposed method is evaluated using a well-known dataset and compared with five state-of-the-art methods. The comparative analysis reveals that the proposed method performs significantly better qualitatively and quantitatively against all the methods.

In semiconductor manufacturing, defect detection is critical to maintain high yield. Currently, computer vision systems used in semiconductor photo lithography still have adopt to digital image processing algorithm, which often occur inspection faults due to sensitivity to external environment. Thus, we intend to handle this problem by means of using Mask R-CNN instead of digital image processing algorithm. Additionally, Mask R-CNN can be trained with image dataset pre-processed by means of the specific designed digital image filter to extract the enhanced feature map of Convolutional Neural Network (CNN). Our approach converged advantage of digital image processing and instance segmentation with deep learning yields more efficient semiconductor photo lithography inspection system than conventional system.

The thermal evaporation source is used to prepare thin films by physical vapor deposition. Materials of metals, organic materials, were tested and explained for thermal evaporation experiments. The developed effusion cell performance depends on the type of deposition material, the size of the crucible, the performance of the reflector, etc. and the proper conditions were found by producing, comparing and analyzing several sets of effusion cell to quantitatively evaluate the performance of the cell. The effusion cell for thermal evaporation source is used to prepare thin films of Ag, Cu, Mg.

Electrochromic (EC) device is an element whose transmittance is changed by electrical energy. Coloring and decoloring states can be easily controlled and thus used in buildings and automobiles for energy saving. There exist several types of EC devices; EC using electrolytes, polymer dispersed liquid crystal (PDLC), and suspended particle device (SPD) using polarized molecules. However, these devices involve solutions such as electrolytes and liquid crystals, limiting their applications in high temperature environments. In this study, we have studied all-solid-state EC device based on Tin(IV) oxide (SnO₂). A coloring phase is achieved when electrons are accumulated in the ultraviolet (UV)-treated SnO₂ layer, whereas a decoloring mode is obtained when electrons are empty there. The UV treatment of SnO₂ layer brings in a number of localized states in the bandgap, which traps electrons near the conduction band. The SnO₂-based EC device shows a transmittance of 70.7% in the decoloring mode and 41% in the coloring mode at a voltage of 2.5 V. We have achieved a transmittance change as large as 29.7% at the wavelength of 550 nm. It also exhibits fast and stable driving characteristics, which have been demonstrated by the cyclic experiments of coloration and decoloration. It has also showed the memory effects induced by the insulating layer of titanium dioxide (TiO₂) and silicone (Si).

A microlens has been fabricated by various methods such as a thermal reflow, hot embossing, diamond milling, etc. However, these methods require a relatively complex process to control the microlens shape. In this work, we report on a simple and cost-effective method to fabricate a cylindrical microlens (CML), which can diffuse light widely. We have employed a slot-die head with the dual plate (a meniscus guide with a protruded μ-tip and a shim with a slit channel) for coating of a narrow stripe using poly(methyl methacrylate) (PMMA). We have shown that the higher the coating gap, the lower the maximum coating speed, which causes an increase in the stripe width and thickness. The coated PMMA stripe has the concave shape. To make it in the shape of a convex microlens, we have applied the thermal reflow method. When the stripe thickness is small, however, its effect is negligible. To increase the stripe thickness, we have increased the number of repeated coating. With this scheme, we have fabricated the CML with the width of 223 ㎛ and the thickness of 7.3 ㎛. Finally, we have demonstrated experimentally that the CML can diffuse light widely, a feature demanded for light extraction efficiency of organic light-emitting diodes (OLEDs) and suppression of moiré patterns in displays.

Silver nanowire (AgNW) random-meshes with high transmittance, low sheet resistance, and high oxidation stability and flexibility were fabricated using solution-based processes. The random-mesh structure was obtained by forming bubbles whose sizes and densities were controlled using a corona treatment of polyethylene terephthalate (PET) substrates. To reduce the sheet resistance of the fabricated AgNW electrode, a washing process using ethanol solution was performed. In addition, nickel (Ni) was coated on AgNW to improve resistance to oxidation. The effects of corona treatment and Ni-coating on the transmittance, sheet resistance, oxidation stability, and flexibility of the AgNW electrodes were investigated.

The demand for stable plasma has been on the rise because of the increased delivery power amount in the chamber for improving productivity, and fast and accurate plasma impedance matching become a crucial performance measure for radio frequency (RF) power system in semiconductor manufacturing equipment. In this paper, the overall impedance matching was understood, and voltage and current values were extracted with voltage - current (VI) probe to measure plasma impedance in real-time. Actual matching data were analyzed to derive calibration coefficient for V and I measurements to understand the characteristics of VI probe, and we demonstrated the tendency of RF impedance matching according to changes in load impedance. This preliminary empirical research can contribute to fast RF matching as well as advanced equipment control for the next level of detailed investigation on embedded system based-RF matching controller.

We studied the electronic and magnetic properties for the Mn-doped chalcopyrite (CH) AlAs, GaAs, and AlGaAs₂ semiconductor by using the first-principles calculations. The chalcopyrite AlGaP₂, AlGaAsP, and AlGaAs₂ compounds have a semiconductor characters with a small band-gap. The interaction between Mn-3d and As-4p states at the Fermi level dominate rather than the other states. The ferromagnetic ordering of dopant Mn with high magnetic moment is induced due to the Mn(3d)-As(4p) strong coupling, which is attributed by the partially filled As-4p bands. The holes are mediated with keeping their 3d-electrons, therefore the ferromagnetic state is stabilized by this double-exchange mechanism. We noted that the ferromagnetic state with high magnetic moment is originated from the hybridized As(4p)-Mn(3d)-As(4p) interaction mediated by the holes-carrier.

The hard mask photo-resist strip inspection system for the semiconductor wafer manufacturing process inspects the position of the circuit pattern formed on the wafer by measuring the distance from the edge of the wafer to the strip processing area. After that, it is an inspection system that enables you to check the process status in real time. Process defects can be significantly reduced by applying a tester that has not been applied to the existing wafer strip process, edge etching process, and wafer ashing process. In addition, it is a technology for localizing semiconductor process inspection equipment that can analyze the outer diameter of the wafer and the state of pattern formation, which can secure process stability and improve wafer edge yield.

This paper is to propose a hybrid memory adaptor using next-generation nonvolatile memory devices such as phase-change memory to improve the performance limitations of OpenStack-based object storage systems. The proposed system aims to improve the performance of the account and container servers for object metadata management. For this, the proposed system consists of locality-based dynamic page buffer, write buffer, and nonvolatile memory modules. Experimental results show that the proposed system improves the hit rate by 5.5% compared to the conventional system.

A new method for array testing of TFT-CD panel with the integrated gate driver circuits is presented. As larger size/high resolution TFT-LCD with the peripheral driver circuits has emerged, one of the important problems for manufacturing is array testing on the panel. This paper describes the technology of detecting defective arrays and optimizing the array testing process. For the effective characterization of pixel array, the pixel storage capability is simulated and measured with voltage imaging system. This technology permits full functional testing during the manufacturing process, enabling fabrication of large TFT-LCD panels with the integrated driver circuits.

The physics-based compact gate leakage current noise models in nanoscale MOSFETs are developed in such a way that the models incorporate important physical effects and are suitable for circuit simulators, including QM (quantum-mechanical) effects. An emphasis on the trap-related parameters of noise models is laid to make the models adaptable to the variations in different process technologies and to make its parameters easily extractable from measured data. With the help of an accurate and generally applicable compact noise models, the compact noise models are successfully implemented into BSIM (Berkeley Short-channel IGFET Model) format. It is shown that the noise models have good agreement with measurements over the frequency, gate-source and drain-source bias ranges.

In this paper, we create scaffold defect classification models using machine learning based data. We extract the characteristic from collected scaffold external images using USB camera. SVM, KNN, MLP algorithm of machine learning was using extracted features. Classification models of three type learned using train dataset. We created scaffold defect classification models using test dataset. We quantified the performance of defect classification models. We have confirmed that the SVM accuracy is 95%. So the best performance model is using SVM.

The comprehensive and physics-based compact noise models for advanced CMOS devices were presented. The models incorporate important physical effects in nanoscale MOSFETs, such as the low frequency correlation effect between the drain and the gate, the trap-related phenomena, and QM (quantum mechanical) effects in the inversion layer. The drain current noise model was improved by including the tunneling assisted-thermally activated process, the realistic trap distribution, the parasitic resistance, and mobility degradation. The expression of correlation coefficient was analytically described, enabling the overall noise performance to be evaluated. With the consideration of QM effects, the comprehensive low frequency noise performance was simulated over the entire bias range.

As a highly conductive and transparent electrode, the optical transmittances of ITO/Ag/ITO were simulated and compared with the experimental results. The simulations are based on the finite-difference time-domain (FDTD) method in solving linear Maxwell equations. In our simulations, the computation domain is set in the XZ-plane with 3D dimension, and a plane wave with variable wavelengths ranging from 250 nm to 850 nm is incident in the z-direction at normal incidence to the ITO/Ag/ITO film surrounded by free-air space. As the results through both simulations and experiments, it was shown that the thickness combinations by the ITO layers of about 40 nm and the Ag layer of about 10 nm could be most suitable conditions as a high conductive transparent electrode having the transmittance similar to that of a single ITO layer.

A majority of car crash is affected by careless driving that causes extensive economic and social costs, as well as injuries and fatalities. Thus, the research of precise crash detection systems is very significant issues in automotive safety. A lot of crash detection algorithms have been developed, but the coverage of these algorithms has been limited to few scenarios. Road scenes and situations need to be considered in order to expand the scope of a collision detection system to include a variety of collision modes. The proposed algorithm effectively handles the x, y, and z axes of the sensor, while considering time and suggests a method suitable for various real worlds. To reduce nuisance and false crash detection events, the algorithm discriminated between driving mode and parking mode. The performance of the suggested algorithm was evaluated under various scenarios, and it successfully discriminated between driving and parking modes, and it adjusted crash detection events depending on the real scenario. The proposed algorithm is expected to efficiently manage the space and lifespan of the storage device by allowing the vehicle's black box system to store only necessary crash event's videos.

Recently, as display and semiconductor devices have been miniaturized and highly integrated, there is a demand for optimization of the structural characteristics of the thin film accordingly. The sputtering device has the advantage of stably obtaining a desired thin film depending on the material selected for the target. However, due to the structural characteristics of the sputtering equipment, the structural characteristics of the film may be different depending on the incidence angle of the sputtering target material to the substrate. In this study, the characteristics of the thin film material according to the scattering angle of the target material and the incidence position of the substrate were studied to find the optimization design rule of the sputtering equipment. To this end, a Si thin film of 1 ㎛ or less was deposited on the Si(100) substrate, and then the microstructure, reflectance, surface roughness, and thin film crystallinity of the thin film formed for each substrate location were investigated. As a result of the study, it was found that as the sputter scattering angle increased and the substrate incident angle decreased, the gap energy along with the surface structure of the thin film increased from 1.47 eV to 1.63 eV, gradually changing to a non-conductive tendency.

In this paper, we propose the non-contact gesture recognition system and algorithm using proximity-based sensors. The system uses four IR receiving photodiode embedded on a single chip and an IR LED for small area. The goal of this paper is to use the proposed algorithm to solve the problem associated with bringing the four IR receivers close to each other and to implement a gesture sensor capable of recognizing eight directional gestures from a distance of 10cm and above. The proposed system was implemented on a FPGA board using Verilog HDL with Android host board. As a result of the implementation, a 2-D swipe gesture of fingers and palms of 3cm and 15cm width was recognized, and a recognition rate of more than 97% was achieved under various conditions. The proposed system is a low-power and non-contact HMI system that recognizes a simple but accurate motion. It can be used as an auxiliary interface to use simple functions such as calls, music, and games for portable devices using batteries.

NAND flash memory is a non-volatile memory that retains stored data even without power supply. Internal memory used as a data storage device and solid-state drive (SSD) is used in portable devices such as smartphones and digital cameras. However, NAND flash memory carries the risk of electric shock, which can cause errors during read/write operations, so use error correction codes to ensure reliability. It efficiently recovers bad block information, which is a defect in NAND flash memory. BBT (Bad Block Table) is configured to manage data to increase stability, and as a result of experimenting with the error correction code algorithm, the bit error rate per page unit of 4Mbytes memory was on average 0ppm, and 100ppm without error correction code. Through the error correction code algorithm, data stability and reliability can be improved.

In this paper, the design method of asynchronous nonvolatile memory module that can efficiently process and store large amounts of data without loss when the power turned off is proposed and implemented. PSRAM, which takes advantage of DRAM and SRAM, was used for data processing, and NAND flash memory was used for data storage and backup. The problem of a lot of signal interference due to the characteristics of memory devices was solved through PCB design using high-density integration technology. In addition, a boost circuit using the super capacitor of 0.47F was designed to supply sufficient power to the system during the time to back up data when the power is off. As a result, an asynchronous nonvolatile memory module was designed and implemented that guarantees reliability and stability and can semi-permanently store data for about 10 years. The proposed method solved the problem of frequent data loss in industrial sites and presented the possibility of commercialization by providing convenience to users and managers.

Autonomous underwater vehicle makes attracts to many researchers. This paper proposes a convolutional neural network (CNN) based fish detection method. Since there are not enough data sets in the process of training, overfitting problem can be occurred in deep learning. To solve the problem, we apply the dropout algorithm to simplify the model. Experimental result showed that the implemented method is promising, and the effectiveness of identification by dropout approach is highly enhanced.

This paper describes the results of a study on the unmanned cleaning robot that performs the cleaning of the floating photovoltaic panels. The robot uses two SSC (Sliding Suction Cup) adsorptive devices to move up and down the slope. First, the forces generated when the robot moves up the slope are mechanically analyzed. The robot was designed and manufactured to operate stably by using the presented results. Next, the robot motion was tested on the inclined panel. It has been proven that robots are well designed and built to clean sloped panels.