• Journal of Korea Multimedia Society
• /
• v.20 no.11
• /
• pp.1750-1758
• /
• 2017

In super-resolution, various methods with Convolutional Neural Network(CNN) have recently been proposed. CNN based methods provide much higher image quality than conventional methods. Especially, VDSR outperforms other CNN based methods in terms of image quality. However, it requires a high computational complexity which prevents real-time processing. In this paper, the method to apply a deconvolution layer to VDSR is proposed to reduce computational complexity. Compared to original VDSR, the proposed method achieves the 4.46 times speed-up and its degradation in image quality is less than -0.1 dB which is negligible.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.15 no.6
• /
• pp.201-208
• /
• 2015

3D image must have depth image for depth information in order for 3D realistic media broadcasting. We used generally belief propagation algorithm to solve probability model. Belief propagation algorithm is operated by message passing between nodes corresponding to each pixel. The high resolution image will be able to precisely represent but that required much computational complexity for 3D representation. We proposed fast stereo matching algorithm using belief propagation with multi-resolution based wavelet or lifting. This method can be shown efficiently computational time at much iterations for accurate disparity map.

• Proceedings of the KSRS Conference
• /
• 1999.11a
• /
• pp.433-438
• /
• 1999

The land use type of Korea is high-density. So, the image classification using coarse resolution satellite image may not provide land cover classification results as good as expected. The purpose of this paper is to compare the result of visual interpretation with that of digital image classification of 20 m resolution SPOT satellite image at Kwangju-eup, Kyunggi-do, Korea. Classes are forest, cultivated field, pasture, water and residential area, which are clearly discriminated in visual interpretation. Maximum likelihood classifier was used for digital image classification. Accuracy assessment was done by comparing each classification result with ground truth data obtained from field checking. The classification result from the visual interpretation presented an total accuracy 9.23 percent higher than that of the digital image classification. This proves the importance of visual interpretation for the area with high density land use like the study site in Korea.

• Earthquakes and Structures
• /
• v.7 no.5
• /
• pp.609-626
• /
• 2014

The capabilities of a high-resolution Digital Image Correlation (DIC) system are presented within the context of deformation measurements of full-scale concrete columns tested under reversed cyclic loading. The system was developed to have very high-resolution such that material strains on the order of the cracking stain of concrete could be measured on the surface of full-scale structural members. The high-resolution DIC system allows the measurement of a wide range of deformations and strains that could only be inferred or assumed previously. The DIC system is able to resolve the full profiles of member curvatures, rotations, plasticity spread, shear deformations, and bar-slip induced rotations. The system allows for automatic and objective measurement of crack widths and other damage indices that are indicative of cumulated damage and required repair time and cost. DIC damage measures contrast prevailing proxy damage indices based on member force-deformation data and subjective damage measures obtained using visual inspection. Data derived from high-resolution DIC systems is shown to be of great use in advancing the state of behavioral knowledge, calibrating behavioral and analytical models, and improving simulation accuracy.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.26 no.3
• /
• pp.234-245
• /
• 2023

The Arena Fragmentation Test(AFT) is designed to analyze warhead performance by measuring fragmentation data. In order to evaluate the results of the AFT, a set of AFT images are captured by high-speed cameras. To detect objects in the AFT image set, ResNet-50 based Faster R-CNN is used as a detection model. However, because of the low resolution of the AFT image set, a detection model has shown low performance. To enhance the performance of the detection model, Super-resolution(SR) methods are used to increase the AFT image set resolution. To this end, The Bicubic method and three SR models: ZSSR, EDSR, and SwinIR are used. The use of SR images results in an increase in the performance of the detection model. While the increase in the number of pixels representing a fragment flame in the AFT images improves the Recall performance of the detection model, the number of pixels representing noise also increases, leading to a slight decreases in Precision performance. Consequently, the F1 score is increased by up to 9 %, demonstrating the effectiveness of SR in enhancing the performance of the detection model.

• Proceedings of the KSRS Conference
• /
• 2005.10a
• /
• pp.213-216
• /
• 2005

A novel hyperspectral imaging spectrometer controlling spatial and spectral resolution individually has been proposed. This imaging spectrometer uses a zoom lens as a telescope and a focusing element. It can change the spatial resolution fixing the spectral resolution or the spectral resolution fixing the spatial resolution. Here, we report the concept of the hyperspectral imaging spectrometer with the novel zooming function and the optical design of a zoom lens as the focusing element. By using lens module and third-order aberration theory, we have presented the initial design of four-group zoom lens with external entrance pupil. And the optimized zoom lens with a focal length of 50 to 150 mm is obtained from the initial design by the optical design software. As a result, the designed zoom lens shows satisfactory performances in wavelength range of 450 to 900 nm as a focusing element in an imaging spectrometer. Furthermore, the collimator lens of the imaging spectrometer is designed through the third-order aberration correction by using an iterative process.

• Journal of the Korean Society of Industry Convergence
• /
• v.21 no.6
• /
• pp.309-316
• /
• 2018

The high-resolution satellite images provided by Kompsat-3A, a multipurpose satellite, have various applications such as digital map generation, 3D image generation, and DEM generation. In order to utilize high-resolution satellite images, the user must create an orthoimage in order to use the image in a suitable manner. The position and the number of the ground reference points affect the accuracy of the orthoimage. In particular, the Kompsat-3A satellite image has a high resolution of about 0.5m, so the difficulty in selecting the ground control points and the accuracy of the selected point will have a great influence on the subsequent application process. Therefore, in this study, we analyzed the influence of the number of ground reference points on the accuracy of the terrestrial satellite images.

• Current Optics and Photonics
• /
• v.2 no.2
• /
• pp.153-164
• /
• 2018

In semiconductor manufacturing, critical dimensions indicate the features of patterns formed by the semiconductor process. The purpose of measuring critical dimensions is to confirm whether patterns are made as intended. The deposition process for an organic light emitting diode (OLED) forms a luminous organic layer on the thin-film transistor electrode. The position of this organic layer greatly affects the luminescent performance of an OLED. Thus, a system for measuring the position of the organic layer from outside of the vacuum chamber in real-time is desired for monitoring the deposition process. Typically, imaging from large stand-off distances results in low spatial resolution because of diffraction blur, and it is difficult to attain an adequate industrial-level measurement. The proposed method offers a new superresolution single-image using a conversion formula between two different optical systems obtained by a deep learning technique. This formula converts an image measured at long distance and with low-resolution optics into one image as if it were measured with high-resolution optics. The performance of this method is evaluated with various samples in terms of spatial resolution and measurement performance.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2000.05a
• /
• pp.233-237
• /
• 2000

Generally, the still image magnification uses .image growing, interpolation in order to get magnificated image. Still image Magnification does not get high-resolution image because, amount of information is not sufficient. In this thesis, we proposed the enhance method of high resolution image magnification. Result of apply proposed method to Lena image, we gained result of enhancement more better than formerly simple technique.

• Journal of Information Processing Systems
• /
• v.13 no.3
• /
• pp.464-475
• /
• 2017

In this paper, we present a new texture image retrieval method which combines color and texture features extracted from images by a set of multi-resolution multi-direction (MRMD) filters. The MRMD filter set chosen is simple and can be separable to low and high frequency information, and provides efficient multi-resolution and multi-direction analysis. The color space used is HSV color space separable to hue, saturation, and value components, which are easily analyzed as showing characteristics similar to the human visual system. This experiment is conducted by comparing precision vs. recall of retrieval and feature vector dimensions. Images for experiments include Corel DB and VisTex DB; Corel_MR DB and VisTex_MR DB, which are transformed from the aforementioned two DBs to have multi-resolution images; and Corel_MD DB and VisTex_MD DB, transformed from the two DBs to have multi-direction images. According to the experimental results, the proposed method improves upon the existing methods in aspects of precision and recall of retrieval, and also reduces feature vector dimensions.