• Korean Journal of Materials Research
• /
• v.32 no.8
• /
• pp.345-353
• /
• 2022

In this study, using deep learning, super-resolution images of transmission electron microscope (TEM) images were generated for nanomaterial analysis. 1169 paired images with 256 × 256 pixels (high resolution: HR) from TEM measurements and 32 × 32 pixels (low resolution: LR) produced using the python module openCV were trained with deep learning models. The TEM images were related to DyVO₄ nanomaterials synthesized by hydrothermal methods. Mean-absolute-error (MAE), peak-signal-to-noise-ratio (PSNR), and structural similarity (SSIM) were used as metrics to evaluate the performance of the models. First, a super-resolution image (SR) was obtained using the traditional interpolation method used in computer vision. In the SR image at low magnification, the shape of the nanomaterial improved. However, the SR images at medium and high magnification failed to show the characteristics of the lattice of the nanomaterials. Second, to obtain a SR image, the deep learning model includes a residual network which reduces the loss of spatial information in the convolutional process of obtaining a feature map. In the process of optimizing the deep learning model, it was confirmed that the performance of the model improved as the number of data increased. In addition, by optimizing the deep learning model using the loss function, including MAE and SSIM at the same time, improved results of the nanomaterial lattice in SR images were achieved at medium and high magnifications. The final proposed deep learning model used four residual blocks to obtain the characteristic map of the low-resolution image, and the super-resolution image was completed using Upsampling2D and the residual block three times.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.40 no.1
• /
• pp.15-22
• /
• 2022

In order to improve the accuracy of the deep learning object detection technique, the effect of magnification rate conditions and seasonal factors on detection accuracy in aerial photographs and drone images was analyzed through experiments. Among the deep learning object detection techniques, Mask R-CNN, which shows fast learning speed and high accuracy, was used to detect the vehicle to be detected in pixel units. Through Seoul's aerial photo service, learning images were captured at different screen magnifications, and the accuracy was analyzed by learning each. According to the experimental results, the higher the magnification level, the higher the mAP average to 60%, 67%, and 75%. When the magnification rates of train and test data of the data set were alternately arranged, low magnification data was arranged as train data, and high magnification data was arranged as test data, showing a difference of more than 20% compared to the opposite case. And in the case of drone images with a seasonal difference with a time difference of 4 months, the results of learning the image data at the same period showed high accuracy with an average of 93%, confirming that seasonal differences also affect learning.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.22 no.9
• /
• pp.1985-1997
• /
• 1997

This paper proposed a regularized iterative image restoration by using method of conjugate gradient. Compared with conventional iterative methods, method of conjugate gradient has a merit to converte toward a solution as a super-linear convergence speed. But because of those properties, there are several artifacts like ringing effects and the partial magnification of the noise in the course of restoring the images that are degraded by a defocusing blur and additive noise. So, we proposed the regularized method of conjugate gradient applying constraints. By applying the projectiong constraint and regularization parameter into that method, it is possible to suppress the magnification of the additive noise. As a experimental results, we showed the superior convergence ratio of the proposed mehtod compared with conventional iterative regularized methods.

• Korean Journal of Optics and Photonics
• /
• v.4 no.3
• /
• pp.252-259
• /
• 1993

A holosymmetric four-mirror system with unit magnification is designed for use in the micro-lithography using a deep ultraviolet wavelength of $0.248 {\mu}m$(KrF excimer laser line). In the holosymmetric system all orders of coma and distortion are zero. By applying this principle to the 4-spherical mirror system, we have obtained only one exact solution for the unit magnification holosymmetric four-spherical mirror system with all zero third order aberrations. For correction of the residual higher order aberrations of the system, aspherization is introduced keeping the holosymmetric properties. We have obtained near diffraction-limited performance for the wavelength of 0.248 pm within N.A. of 0.33 and image field diameter of 7.6 mm.

• Journal of the Korea Society of Computer and Information
• /
• v.21 no.12
• /
• pp.19-26
• /
• 2016

In this paper, we propose an efficient design method for a thermal camera with continuous zoom based on the research and manufacturing experience of the thermal camera. In addition, it is divided into system design method, optical design method, mechanical design method, and electronic design method. First, we propose an effective NUC compensation method and a lens-specific sensitivity design method in terms of system. Second, we propose a zoom trajectory design method considering the temperature effect on the optical aspect. Third, it suggests the minimization of optical axis shaking between magnification conversion in terms of mechanism. Finally, we propose a lens-specific temperature compensation method and a speed conversion algorithm according to the zoom interval as an electronic aspect.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.31 no.12C
• /
• pp.1184-1193
• /
• 2006

Image interpolation has been widely used and studied in the various fields of image processing. There are many approaches of varying complexity and robustness. In this paper, a new distance weight is proposed for the conventional linear interpolation. In comparison with the conventional linear weight, the new distance weight uses a quadratic or cubic polynomial equation to reflect that the interpolated value should be influenced more by the value of closer pixels in an input image. In this paper, the new adaptive linear (NAL) interpolation, which considers patterns near the interpolated value, is also proposed. This algorithm requires a pattern weight, which is used to determine the ratio of reflection on local patterns, to obtain an interpolated image that exhibits better quality at various magnification factors (MF). In the computer simulation, not only did the NAL interpolation exhibit much lower computational complexity than conventional bicubic interpolation, it also improved peak signal-to-noise ratios (PSNR).

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 1996.11a
• /
• pp.490-495
• /
• 1996

Though the increment of using computer vision system in modern industry, there are lots of difficulties to measure precisely because of measurement error distortion phenomenon. Among these reasons, the distortion of edge is dominant reason which is occurred by the blurred image. The blurred image is happened when camera can not discriminate its precise focus. To calibrate and generalize distortion phenomenon is important. Thus, we must fix the discrimination criteria which is collected by image recognition of precise focus. Also, radial distortion causes an inward or outward displacement of a given image point from its ideal location. This type of distortion is mainly caused by flawed radial curvature curve of the elements. Thus, we were analyzed the distortion in terms of the changed with lens magnification.

• Applied Microscopy
• /
• v.2 no.1
• /
• pp.39-46
• /
• 1972

There are many kinds of microscopes suitable for general studies; optical microscopes(OM), conventional transmission electron microscopes (TEM), and scanning electron microscopes(SEM). The optical microscopes and the conventional transmission electron microscopes are very familiar. The images of these microscopes are directly formed on an image plane with one or more image forming lenses. On the other hand, the image of the scanning electron microscope is formed on a fluorescent screen of a cathode ray tube using a scanning system similar to television technique. In this paper, the features and some applications of the scanning electron microscope will be discussed briefly. The recently available scanning electron microscope, combining a resolution of about $200{\AA}$ with great depth of field, is favorable when compared to the replica technique. It avoids the problem of specimen damage and the introduction of artifacts. In addition, it permits the examination of many samples that can not be replicated, and provides a broader range of information. The scanning electron microscope has found application in diverse fields of study including biology, chemistry, materials science, semiconductor technology, and many others. In scanning electron microscopy, the secondary electron method. the backscattererd electron method, and the electromotive force method are most widely used, and the transmitted electron method will become more useful. Change-over of magnification can be easily done by controlling the scanning width of the electron probe. It is possible. to continuously vary the magnification over the range from 100 times to 1.00,000 times without readjustment of focusing. Conclusion: With the development of a scanning. electron microscope, it is now possible to observe almost all-information produced through interactions between substances and electrons in the form of image. When the probe is properly focused on the specimen, changing magnification of specimen orientation does not require any change in focus. This is quite different from the conventional transmission electron microscope. It is worthwhile to note that the typical probe currents of $10^{-10}$ to $10^{-12}\;{\AA}$ are for below the $10^{-5}$ to $10^{-7}\;{\AA}$ of a conventional. transmission microscope. This reduces specimen contamination and specimen damage due to heatings. Outstanding features of the scanning electron microscope include the 'stereoscopic observation of a bulky or fiber specimen in high resolution' and 'observation of potential distribution and electromotive force in semiconductor devices'.

• Journal of the Korean Society of Manufacturing Technology Engineers
• /
• v.19 no.4
• /
• pp.562-569
• /
• 2010

This paper analyzes the problems that occurred in the magnification process for a fine input image and investigates a method to improve the blurring of magnified image. This paper applies a curve interpolation algorithm in CAD/CAM for the same test images with the existing image algorithm in order to improve the blurring of magnified image. As a result, the nearest neighbor interpolation, which is the most frequently applied algorithm for the existing image interpolation algorithm, shows that the identification of a magnified image is not possible. Therefore, this study examines an interpolation of gray-level data by applying a low-pass spatial filter and verifies that a bilinear interpolation presents a lack of property that accentuates the boundary of the image where the image is largely changed. The periodic B-spline interpolation algorithm used for curve interpolation in CAD/CAM can remove the blurring but shows a problem of obscuration, and the Ferguson' curve interpolation algorithm shows a more sharpened image than that of the periodic B-spline algorithm. For the future study, hereafter, this study will develop an interpolation algorithm that has an excellent improvement for the boundary of the image and continuous and flexible property by using the NURBS, Ferguson' complex surface, and Bezier surface used in CAD/CAM engineering based on the results of this study.

• Journal of the Korea society of information convergence
• /
• v.1 no.1
• /
• pp.17-24
• /
• 2008

Video camera play very important roles for preventing many kinds of crimes and resolving those crime affairs. But in the case of recording image of a specific person far from the CCTV, the original image needs to be enlarged and recovered in order to identify the person more obviously. Interpolation is usually used for the enlargement and recovery of the image in this case. However, it has a certain limitation. As the magnification of enlargement is getting bigger, the quality of the original image can be worse. This paper uses FOP(Facial Definition Parameter) proposed by the MPEG-4 SNHC FBA group and introduces a new algorithm that uses face outline information of the original image based on the FOP, which makes it possible to recover better than the known methods until now.