• Journal of Information Processing Systems
• /
• v.14 no.1
• /
• pp.205-217
• /
• 2018

Deep learning using convolutional neural networks (CNNs) is being studied in various fields of image recognition and these studies show excellent performance. In this paper, we compare the performance of CNN architectures, KCR-AlexNet and KCR-GoogLeNet. The experimental data used in this paper is obtained from PHD08, a large-scale Korean character database. It has 2,187 samples of each Korean character with 2,350 Korean character classes for a total of 5,139,450 data samples. In the training results, KCR-AlexNet showed an accuracy of over 98% for the top-1 test and KCR-GoogLeNet showed an accuracy of over 99% for the top-1 test after the final training iteration. We made an additional Korean character dataset with fonts that were not in PHD08 to compare the classification success rate with commercial optical character recognition (OCR) programs and ensure the objectivity of the experiment. While the commercial OCR programs showed 66.95% to 83.16% classification success rates, KCR-AlexNet and KCR-GoogLeNet showed average classification success rates of 90.12% and 89.14%, respectively, which are higher than the commercial OCR programs' rates. Considering the time factor, KCR-AlexNet was faster than KCR-GoogLeNet when they were trained using PHD08; otherwise, KCR-GoogLeNet had a faster classification speed.

• Proceedings of the Korean Institute of Electrical and Electronic Material Engineers Conference
• /
• 2007.06a
• /
• pp.398-398
• /
• 2007

Most recently, the Liquid Crystal (LC) aligning capabilities achieved by ion beam exposure on the diamond-like carbon (DLC) thin film layer have been successfully studied. The DLC thin films have a high mechanical hardness, a high electrical resistance, optical transparency and chemical inertness. Nitrogen doped Diamond Like Carbon (NDLC) thin films exhibit properties similar to those of the DLC films and better thermal stability than the DLC films because C:N bonding in the NDLC film is stronger against thermal stress than C:H bonding in the DLC thin films. Moreover, our research group has already studied ion beam alignment method using the NDLC thin films. The nematic liquid crystal (NLC) alignment effects treated on the SiNx thin film layers using ion beam irradiation for three kinds of N rations was successfully studied for the first time. The SiNx thin film was deposited by plasma-enhanced chemical vapor deposition (PECVD) and used three kinds of N rations. In order to characterize the films, the atomic force microscopy (AFM) image was observed. The good LC aligning capabilities treated on the SiNx thin film with ion beam exposure for all N rations can be achieved. The low pretilt angles for a NLC treated on the SiNx thin film with ion beam irradiation were measure.

• Korean Journal of Optics and Photonics
• /
• v.3 no.2
• /
• pp.77-85
• /
• 1992

The design of a flat-field XUV spectrograph is optimized for a high throughput, aberration-corrected spectral image in the wavelength region of 50-300 $\AA$ The varied-line spacing (VLS) concave grating theory for an XUV spectrograph with a toroidal mirror in front of an entrance slit is derived. Since the derived theory includes the arbitrary shaped source, it is able to correct the limit of the simple optimization theory which considers only a point source at the center of the entrance slit. The reflection matrix at the toroidal mirror and the diffraction matrix at VLS grating are derived and compared with those of a holographic grating. The absolute energy efficiency of a flat-field spectrograph is calculated by considering the reflectivities of the toroidal mirror and the Au coated concave grating and the grating efficiency. The alignment sensitivity of the toroidal mirror and the concave grating is investigated, and the method to achieve the best imaging of XUV spectrum is discussed. The calculated resolving power of the flat-field XUV spectrograph is more than 4000 in the aberration-corrected wavelength range. The focused spot size at the dispersion plane is less than $20\mu \textrm m\times \mu \textrm m$at the wavelength 100$\AA$ It is shown that a high throughput characteristic can also be achieved through a careful adjustment of alignment parameters.

• Journal of Korean Society of Occupational and Environmental Hygiene
• /
• v.28 no.3
• /
• pp.273-282
• /
• 2018

Objectives: Dustiness of nanomaterials is considered as exposure index of essential material. Research on dustiness of nanomaterial is needed to control exposure in workplaces. Method: Dustiness measurement using vortex shaker were installed in the laboratory. Nanomaterials, 1 g, was put in the glass test tube and shaked using vortex shaker. Aerosol dispersed was measured using scanning mobility particle sizer(SMPS) and optical particle counter(OPC). Mass concentration using PVC filter and cassette was measured and TEM grid sampling was conducted. Total particle concentration and size distribution were calculated. Image and chemical composition of particles in the air were observed using transmission electron microscopy and energy dispersive X-ray spectrometer. Eleven different test nanomaterials were used in the study. Results: Rank of mass concentration and particle number concentration were coincided in most cases. Rank of nanomateirals with low concentration were not coincided. Two types of fumed silica had the highest mass concentration and particle number concentration. Indium tin oxide, a mixture of indium oxide and tin oxide, had high mass concentration and particle number concentration. Indium oxide had very low mass concentration and particle number concentration. Agglomeration of nanoparticles in the air were observed in TEM analysis and size distribution. In this study, mass concentration and particle number concentration were coincided and two index can be used together. The range of dustiness in particle number concentration were too wide to measure in one method. Conclusion: Particle number concentration ranged from low concentration to high concentration depend on type of nanomaterial, and varied by preparation and amount of nanomaterial used. Further study is needed to measure dustiness of all nanomaterial as one reference method.

• Journal of Korea Multimedia Society
• /
• v.9 no.5
• /
• pp.585-594
• /
• 2006

Due to the size limit, the voice coil motor (VCM) is adopted in most of the mobile phone camera to control auto-focus instead of step motor. The optical system using the VCM has the property that the focus values are varying even though the same current is induced. It means that an error of the lens position was taken placed due to the characteristics of the VCM. In this paper, a algorithm was proposed to compensate the lens position error using the step size and the search count of each stage. In the proposed algorithm -7 step middle searching stage is inserted the conventional searching algorithm for the fast auto-focus searching and the final searing step size was set to +1 for the precise focus control, respectively. In the experimental results, the focus values was found more fast in the proposed algorithm than the conventional. And more the image quality by the proposed algorithm was superior to that of the conventional.

• Proceedings of the KWS Conference
• /
• 2009.11a
• /
• pp.95-95
• /
• 2009

Zinc coatings provide the most effective and economical way of protecting steel against corrosion. There are three types of galvanizing lines typically used in production line in galvanizing industries,Galvanize (GI) coating (Zn-0.1-0.3%Al), Galfan coating (Zn-5%Al), Galvalume(GL) coating (45%Zn-Al). In continuous Galvanizing lines, the immersed bath hardware (e.g. bearings, sink, stabilizer, and corrector rolls, and also support roll arms and snout tip) are subjected to corrosion and wear failure. Understanding the reaction of these materials with the molten Zn alloy is becomes scientific and commercial interest. To investigate the reaction with molten Zn alloys, static immersion test performed for 4, 8, 16, and 24 Hr. Two different baths used for the static immersion, which are molten Zn and molten Zn-55%Al. Microstructures characterization of each of the materials and intermetallic layer formed in the reaction zone was performed using optical microscope, SEM and EDS. The thickness of the reaction layer is examined using image analysis to determine the kinetics of the reaction. The phase dominated by two distinct phase which are eutectic carbide and matrix. The morphology of the intermetallic phase formed by molten Zn is discrete phase showing high dissolution of the material, and the intermetallic phase formed by Zn-55wt%Al is continuous. Aluminum reacts readily with the materials compare to Zinc, forming iron aluminide intermetallic layer ($Fe_2Al_5$) at the interface and leaving zinc behind.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.11 no.2
• /
• pp.337-343
• /
• 2007

Since three-dimensional (3D) images reconstructed in interval imaging technique are related to the resolution of elemental images, there has been a problem that ray information of elemental images increases largely in order to obtain high-resolution 3D images. In this paper, to overcome this problem, a new unidirectional integral imaging based on pinhole model is proposed. Proposed method provides a new type of unidirectional elemental images, which are simply obtained by magnifying single horizontal pixel line of each elemental image to the vertical size of lenslet using ray analysis based on pinhole model and used to display 3D images. In proposed method, reduction effect of the ray information of elemental images can be obtained by scarifying vortical parallax. Feasibility of the proposed scheme is experimentally demonstrated and its results are presented.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.11 no.5
• /
• pp.991-996
• /
• 2007

Recently, a computational reconstruction method using an integral imaging technique, which is a promise three-dimensional display technique, has been actively researched. This method is that 3-D images can be digitally reconstructed at the required output planes by superposition of all of the inversely enlarged elemental images by using a hypothetical pinhole array model. However, the conventional method mostly yields reconstructed images having a low-resolution, because there are some intensity irregularities with a grid structure at the reconstructed mage plane by using square-type elemental images. In this paper, to overcome this problem, we propose a novel computational integral imaging reconstruction (CIIR) method using round-type mapping model. Proposed CIIR method can overcome problems of non-uniformly reconstructed images caused from the conventional method and improve the resolution of 3-D images. To show the usefulness of the proposed method, both computational experiment and optical experiment are carried out and their results are presented.

• Korean Journal of Remote Sensing
• /
• v.13 no.2
• /
• pp.85-98
• /
• 1997

A geometric model of pushbroom-type linear CCD camera images is proposed in this paper. At present, this type of cameras are used for obtaining almost all kinds of high-resolution optical images from satellites. The proposed geometric model includes not only a forward transformation which is much more efficient. An inverse transformation function cannot be derived analytically in a closed form because the focal point of an image varies with time. In this paper, therefore, an iterative algorithm in which a focal point os converged to a given pixel position is proposed. Although the proposed model can be applied to any pushbroom-type linear CCD camera images, the geometric model of the high-resolution multi-spectral camera on-board KITSAT-3 is used in this paper as an example. The flight model of KITSAT-3 is in development currently and it is due to be launched late 1998.

• Korean Journal of Remote Sensing
• /
• v.19 no.6
• /
• pp.411-419
• /
• 2003

An area corresponding to the spatial resolution of optical remote sensor imagery often includes more than one pure surface material. In such case, a pixel value represents a mixture of spectral reflectance of several materials within it. This study attempts to apply the spectral mixture analysis on forest and to evaluate the information content of endmember fractions resulted from the spectral unmixing. Landsat-7 ETM+ image obtained over the study area in the Kwangneung Experimental Forest was initially geo-referenced and radiometrically corrected to reduce the atmospheric and topographic attenuations. Linear mixture model was applied to separate each pixel by the fraction of six endmember: deciduous, coniferous, soil, built-up, shadow, and rice/grass. The fractional values of six endmember could be used to separate forest cover in more detailed spatial scale. In addition, the soil fraction can be further used to extract the information related to the canopy closure. We also found that the shadow effect is more distinctive at coniferous stands.