• Journal of Multimedia Information System
• /
• 제8권2호
• /
• pp.101-110
• /
• 2021

In computer vision, single-image super resolution has been an area of research for a significant period. Traditional techniques involve interpolation-based methods such as Nearest-neighbor, Bilinear, and Bicubic for image restoration. Although implementations of convolutional neural networks have provided outstanding results in recent years, efficiency and single model multi-scalability have been its challenges. Furthermore, previous works haven't placed enough emphasis on real-number scalability. Interpolation-based techniques, however, have no limit in terms of scalability as they are able to upscale images to any desired size. In this paper, we propose a convolutional neural network possessing the advantages of the interpolation-based techniques, which is also efficient, deeming it suitable in practical implementations. It consists of convolutional layers applied on the low-resolution space, post-up-sampling along the end hidden layers, and additional layers on high-resolution space. Up-sampling is applied on a multiple channeled feature map via bicubic interpolation using a single model. Experiments on architectural structure, layer reduction, and real-number scale training are executed with results proving efficient amongst multi-scale learning (including scale multi-path-learning) based models.

• 대한임베디드공학회논문지
• /
• 제3권1호
• /
• pp.8-18
• /
• 2008

In this paper, we propose an efficient algorithm using wavelet transform for still image database retrieval. Especially, it uses only the lowest frequency sub-band in multi-level wavelet transform so that a retrieval system uses a smaller quantity of memory and takes a faster processing time. We extract different textured features, statistical information such as mean, variance and histogram, from low frequency sub-band. Then we measure the distances between the query image and the images in a database in terms of these features. To obtain good retrieval performance, we use the first feature (mean and variance of wavelet coefficients) to filter out most of the unlikely images. The rest of the images are considered to be candidate images. Then we apply the second feature (histogram of wavelet coefficient) to rank all the candidate images. To evaluate the algorithm, we create various distorted image databases using MIT VisTex texture images and PICS natural images. Through simulations, we demonstrate that our method can achieve performance satisfactorily in terms of the retrieval accuracy as well as the both memory requirement and computational complexity. Therefore it is expected to provide good retrieval solution for JPEG-2000 using wavelet transform.

• 산업경영시스템학회지
• /
• 제46권3호
• /
• pp.268-274
• /
• 2023

Image mosaicking is one of the basic and important technologies in the field of application using images. The key of image mosaicking is to extract seamlines from a joint image. The method proposed in this paper for image mosaicking is as follows. The feature points of the images to be joined are extracted and the joining form between the two images is identified. A reference position for detection the seamlines were selected according to the joint form, and an image pyramid was created for efficient image processing. The outlines of the image including buildings and roads are extracted from the overlapping area with low resolution, and the seamlines are determined by considering the components of the outlines. Based on this, the seamlines in the high-resolution image was re-searched and finally the seamline for image mosaicking was determined. In addition, in order to minimize color distortion of the image with the determined seamline, a method of improving the quality of the mosaic image by fine correction of the mosaic area was applied. It was confirmed that the quality of the seamline extraction results applying the method proposed was reasonable.

• 한국정보통신학회논문지
• /
• 제24권12호
• /
• pp.1604-1611
• /
• 2020

마이크로렌즈 어레이 기반의 카메라로 촬영된 라이트필드 영상은 낮은 공간해상도 및 각해상도로 인하여 실제 사용하기에는 많은 제약이 따른다. 고해상도의 공간해상도 영상은 최근 많이 연구되고 있는 단일 영상 초해상도 기법으로 쉽게 얻을 수 있으나 고해상도의 각해상도 영상은 영상사이에 내재된 시점차 정보를 이용하는 과정에서 왜곡이 발생하여 좋은 품질의 각해상도 영상을 얻기 힘든 문제가 있다. 본 논문에서는 영상 사이에 내재된 시점차 정보를 효과적으로 추출하기 위해서 팽창 합성곱 신경망을 이용하여 초기 특징맵을 추출하고 잔차 신경망으로 새로운 시점 영상을 생성하는 라이트 필드 각 초해상도 영상 기법을 제안한다. 제안하는 네트워크는 기존의 각 초해상도 네트워크와 비교하여 PSNR 및 주관적 화질 비교에서 우수한 성능을 보였다.

• 대한원격탐사학회지
• /
• 제36권3호
• /
• pp.449-460
• /
• 2020

초해상화 기법은 저해상도 영상을 고해상도 영상으로 변환하는 기법이다. 최근에는 딥러닝 기술을 활용한 초해상화 방법이 주류를 이루고 있으며, 원격 탐사 분야에서도 이를 응용한 연구가 증가하고 있다. 본 연구에서는 위성 영상의 4배 해상도 향상을 위하여 deep back-projection network (DBPN) 네트워크에 기반한 초해상화 기법을 제안하였다. 또한, 복원된 영상의 디테일 및 윤곽선 부분에서의 고품질 영상 획득을 위해 윤곽선 손실 함수를 제안하고, 효과적이고 안정적인 학습을 위하여 Wasserstein distance 손실 함수를 사용한 GAN 기법을 적용하였다. 또한, 자연스러운 저해상도 훈련 영상을 획득하기 위한 detail preserving image downscaling (DPID) 기법을 적용하였다. 마지막으로 전정 영상의 특징을 추출하여 훈련의 마지막 단계에 적용 시킴으로써 출력 영상의 세부적인 특징을 효과적으로 생성하였다. 그 결과 실험에 사용된 WorldView-3 영상 및 KOMPSAT-2 영상에서 해상도 향상 효과를 확인하였고, 다른 초해상화 모델에 대비하여 윤곽선 보존력이나 영상의 선명도가 향상 되었음을 확인하였다

• 한국정보통신학회논문지
• /
• 제26권12호
• /
• pp.1816-1825
• /
• 2022

본 논문에서는 다양한 해상도의 카메라를 사용하는 감시 카메라 네트워크에서 각 사람을 추적하는 새로운 접근 방식을 제안한다. 다수의 비겹침 카메라 상에서 사람 추적 시 기존에 사용되던 사람 특징 정보는 다양한 카메라 시야 조건에 쉽게 영향을 받는다. 이러한 한계를 극복하기 위해 제안하는 시스템은 외모 정보와 함께 얼굴 정보를 활용한다. 일반적으로 감시 카메라로 촬영하는 사람 영상은 해상도가 낮은 경우가 많기 때문에 추적을 용이하게 하기 위해 저해상도 얼굴에서도 유용한 특징을 추출할 수 있어야 한다. 제안하는 추적 방식에서 사람 얼굴 특징을 추출하기 위해 탐지된 얼굴을 정면화한 후 텍스쳐 기반의 특징을 추출한다. 또한 감시 카메라에 포착된 얼굴의 크기가 매우 작은 경우 얼굴을 확대하는 초해상도 기법도 함께 활용한다. 공개된 데이터셋인 Dana36을 이용하여 수행한 실험결과를 통해 제안된 알고리즘의 우수한 성능을 보여준다.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 2005년도 춘계학술대회 논문집
• /
• pp.1871-1874
• /
• 2005

The direct write FIB technology has several advantages over contemporary micro-machining technology, including better feature resolution with low lateral scattering and capability of maskless fabrication. Therefore, the application of FIB technology in micro fabrication has become increasingly popular. In recent model of FIB, however the feeding system has been a very coarse resolution of about a few ${\mu}m$. Our research is the development of nano stage of 200mm strokes and 10nm resolutions. Also, this stage should be effectively operating in ultra high vacuum of about $1x10^{-7}$ torr. This paper presents the discussion and results of CAE of the 2 axes stages. we have estimated the stable static and dynamic characteristics for dual servo system. Therefore the 2 axes stages developed and future work are introduced at the end of the paper.

• Journal of Information Display
• /
• 제8권4호
• /
• pp.27-31
• /
• 2007

In this paper, we present an efficient tiled stereo display system for tangible meeting. For tangible meeting, it is important to provide immersive display with high resolution image to cover up the field of view and provide to the local user the same environment as that of remote site. To achieve these, a high resolution image needs to be transmitted for reconstruction of remote world, and it should be displayed using a tiled display. However, it is hard to transmit high resolution image in real time due to the limit of network bandwidth, and so we receive multiple images and reconstruct a remote world with received images in advance. Then, we update only a specific area where remote user exists by receiving low resolution image in realtime. We synthesize the transmitted image to the existing environmental map of remote world and display it as a stereo image. For this, we developed a new system which supports GPU based real time warping and blending, automatic feature extraction using machine vision technique.

• Smart Structures and Systems
• /
• 제5권4호
• /
• pp.345-355
• /
• 2009

This article describes the development of a fiber optic accelerometer based on Fiber Bragg Gratings (FBG). The accelerometer utilizes the stiffness of the optical fiber and a lumped mass in the design. Acceleration is measured by the FBG in response to the vibration of the fiber optic mass system. The wavelength shift of FBG is proportional to the change in acceleration, and the gauge factor pertains to the shift in wavelength as a function of acceleration. Low frequency version of the accelerometer was developed for applications in monitoring bridges. The accelerometer was first evaluated in laboratory settings and then employed in a demonstration project for condition assessment of a bridge. Laboratory experiments involved evaluation of the sensitivity and resolution of measurements under a series of low frequency low amplitude conditions. The main feature of this accelerometer is single channel multiplexing capability rendering the system highly practical for application in condition assessment of bridges. This feature of the accelerometer was evaluated by using the system during ambient vibration tests of a bridge. The Frequency Domain Decomposition method was employed to identify the mode shapes and natural frequencies of the bridge. Results were compared with the data acquired from the conventional accelerometers.

• 한국공작기계학회:학술대회논문집
• /
• 한국공작기계학회 2004년도 춘계학술대회 논문집
• /
• pp.472-477
• /
• 2004

Miniaturization is the central theme in modern fabrication technology. Many of the components used in modem products are becoming smaller and smaller. The direct write FIB technology has several advantages over contemporary micromachining technology, including better feature resolution with low lateral scattering and capability of mastless fabrication. Therefore, the application of focused ion beam(FIB) technology in micro fabrication has become increasingly popular. In recent model of FIB, however the feeding system has been a very coarse resolution of about a few ${\mu}{\textrm}{m}$. It is not unsuitable to the sputtering and the deposition to make the high-precision structure in micro or macro scale. Our research is the development of nano stage of 200mm strokes and l0nm resolutions. Also, this stage should be effectively operating in ultra high vacuum of about 1$\times$10$^{-5}$ pa. This paper presents the concept of nano stages and the discussion of the material treatment for ultra tush vacuum.