• 디지털정책학회지
• /
• 제1권2호
• /
• pp.53-59
• /
• 2022

본 논문에서는 단일 영상을 기반으로 초해상도를 생성하는 과정에서 계층 구조를 거쳐 추출된 보조 경계 특징을 이용한 방법을 제안하였다. 초해상도의 품질을 향상시키기 위해서는 영상 내 경계 영역을 선명하게 표현하면서도 각 영역의 형태를 명확하게 구분하여야 한다. 제안하는 방법은 초해상도 과정에서 품질을 결정하는 중요한 요인인 경계 영역을 입력 영상의 구조적 형태를 유지하면서 개선된 초해상도 결과를 생성하기 위해 딥러닝 기반의 초해상도 방법에서 영상의 경계 영역 정보를 보조적으로 활용하는 구조를 사용하였다. 딥러닝 기반의 초해상도를 수행하기 위한 그룹 컨볼루션 구조에 더해 보조 경계 추출을 위한 고주파 대역의 정보를 기반으로 별도의 계층적 구조의 경계 누적 추출 과정을 수행하여 이를 보조 특징으로써 활용하는 방법을 제안하였다. 실험 결과 기존 초해상도 대비 PSNR과 SSIM에서 약 1%의 성능 향상을 보였다.

• 대한원격탐사학회지
• /
• 제39권4호
• /
• pp.459-466
• /
• 2023

In the last decade, artificial intelligence's dramatic advancement with the development of various deep learning techniques has significantly contributed to remote sensing fields and satellite image applications. Among many prominent areas, super-resolution research has seen substantial growth with the release of several benchmark datasets and the rise of generative adversarial network-based studies. However, most previously published remote sensing benchmark datasets represent spatial resolution within approximately 10 meters, imposing limitations when directly applying for super-resolution of small objects with cm unit spatial resolution. Furthermore, if the dataset lacks a global spatial distribution and is specialized in particular land covers, the consequent lack of feature diversity can directly impact the quantitative performance and prevent the formation of robust foundation models. To overcome these issues, this paper proposes a method to generate benchmark datasets by simulating the modulation transfer functions of the sensor. The proposed approach leverages the simulation method with a solid theoretical foundation, notably recognized in image fusion. Additionally, the generated benchmark dataset is applied to state-of-the-art super-resolution base models for quantitative and visual analysis and discusses the shortcomings of the existing datasets. Through these efforts, we anticipate that the proposed benchmark dataset will facilitate various super-resolution research shortly in Korea.

• 정보저장시스템학회논문집
• /
• 제1권2호
• /
• pp.132-136
• /
• 2005

It is widely accepted that the reading mechanism of Super-RENS(super-resolution near field structure) and Super-ROM(super-resolution read only memory) is closely related with non-linear temperature dependent material properties such as refractive indices, phase change. Furthermore, the dynamic change of the temperature distribution also an essential part of reading mechanism of Super-RENS/ROM. Therefore, the knowledge of the temperature distribution as a function a time is one of the important keys to reveal the physics of reading mechanism in Super-RENS/ROM. We calculated time-dependent temperature distribution in a 3-dimensional Super-ROM disk structure when moving laser beam is irradiated. With a help of commercial software FEMLAB which employed finite element method, we simulated the temperature distribution of ROM structure whose pit diameter is 120-nm with 50-nm depth. Energy absorption by moving laser irradiation, time variations of heat transfer processes, heat fluxes, heat transfer ratios, and temperature distributions of the complicate 3-dimensional ROM structure have been obtained.

• ETRI Journal
• /
• 제32권4호
• /
• pp.596-602
• /
• 2010

This paper proposes a computationally efficient learning-based super-resolution algorithm using k-means clustering. Conventional learning-based super-resolution requires a huge dictionary for reliable performance, which brings about a tremendous memory cost as well as a burdensome matching computation. In order to overcome this problem, the proposed algorithm significantly reduces the size of the trained dictionary by properly clustering similar patches at the learning phase. Experimental results show that the proposed algorithm provides superior visual quality to the conventional algorithms, while needing much less computational complexity.

• Current Optics and Photonics
• /
• 제6권6호
• /
• pp.550-564
• /
• 2022

Optical microscopy is a useful tool for study in the biological sciences. With an optical microscope, we can observe the micro world of life such as tissues, cells, and proteins. A fluorescent dye or a fluorescent protein provides an opportunity to mark a specific target in the crowd of biological samples, so that an image of a specific target can be observed by an optical microscope. The optical microscope, however, is constrained in resolution due to diffraction limit. Super-resolution microscopy made a breakthrough with this diffraction limit. Using a super-resolution microscope, many biomolecules are observed beyond the diffraction limit in cells. In the case of volumetric imaging, the super-resolution techniques are only applied to a limited area due to long imaging time, multiple scattering of photons, and sample-induced aberration in deep tissue. In this article, we review recent advances in super-resolution microscopy for volumetric imaging. The super-resolution techniques have been integrated with various modalities, such as a line-scan confocal microscope, a spinning disk confocal microscope, a light sheet microscope, and point spread function engineering. Super-resolution microscopy combined with adaptive optics by compensating for wave distortions is a promising method for deep tissue imaging and biomedical applications.

• 한국멀티미디어학회논문지
• /
• 제22권11호
• /
• pp.1242-1250
• /
• 2019

Super resolution technique aims to convert a low-resolution image with coarse details to a corresponding high-resolution image with refined details. In the past decades, the performance is greatly improved due to progress of deep learning models. However, universal solution for various objects is a still challenging issue. We observe that learning super resolution with a general dataset has poor performance on faces. In this paper, we propose a super resolution fusion scheme that works well for both general- and face datasets to achieve more universal solution. In addition, object-specific feature extractor is employed for better reconstruction performance. In our experiments, we compare our fusion image and super-resolved images from one- of the state-of-the-art deep learning models trained with DIV2K and FFHQ datasets. Quantitative and qualitative evaluates show that our fusion scheme successfully works well for both datasets. We expect our fusion scheme to be effective on other objects with poor performance and this will lead to universal solutions.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2005년도 추계종합학술대회
• /
• pp.481-484
• /
• 2005

Inherent opportunities on research for restoring high resolution image from low resolution images are increasing in these days. Super resolution image reconstruction is the process of combining multiple low resolution images to form a higher resolution one. To achieve super resolution reconstruction, proper observation model which is based on subpixel shift information is required. In this context, the importance of the subpixel registration cannot be estimated because subpixel shift information cannot be obtained from original image. This paper presents a regularized adaptive super resolution reconstruction method based on phase correlated subpixel registration, where the Constrained Least Squares(CLS) Restoration is adopted as a post process.

• 한국지능시스템학회논문지
• /
• 제17권1호
• /
• pp.72-79
• /
• 2007

저해상도 영상 정보들 이용하여 고해상도 영상으로 재구성하는 새로운 고해상도 복원 알고리즘을 제안한다. 제안된 고해상도 복원 알고리즘은 super 해상도 이론을 바탕으로 구성되며, super 해상도는 정합과 복원의 순차적인 단계로 구성되어있다. 본 논문에서는 다해상도 분해를 통한 웨이브렛 기저와 하위픽셀이동을 통한 정합으로 많은 데이터 처리량과 잡음을 줄여 주요정보 유지와 에러율 개선하였다. 또한 복원단계에서는 퍼지 웨이브렛 B-스플라인 보간법을 이용하여 블러링과 블록화 현상이 없는 부드러운 영상과 해상도를 얻음을 확인하였다.

• 대한조선학회논문집
• /
• 제59권2호
• /
• pp.72-79
• /
• 2022

Deep-learning-based Super-Resolution (SR) methods were evaluated to reconstruct pressure fields with a high resolution from low-resolution images taken from a coarse grid simulation. In addition to a canonical SRCNN(super-resolution convolutional neural network) model, two modified models from SRCNN, adding an activation function (ReLU or Sigmoid function) to the output layer, were considered in the present study. High resolution images obtained by three models were more vivid and reliable qualitatively, compared with a conventional super-resolution method of bicubic interpolation. A quantitative comparison of statistical similarity showed that SRCNN model with Sigmoid function achieved best performance with less dependency on original resolution of input images.

• 한국정보통신학회논문지
• /
• 제17권7호
• /
• pp.1672-1686
• /
• 2013

본 논문에서는 이산 웨이블릿 변환(discrete wavelet transform: DWT)을 이용한 단일영상 기반의 초고해상도 기법(super-resolution)과, 복수영상 기반의 초고해상도 기법을 제시하고 두 기법을 혼합한 새로운 초고해상도 기법 기법을 제안한다. 기존의 단일 영상 기반의 초고해상도 기법의 경우 처리 시간이 빠르다는 장점이 있으나 영상 보간 시 사용할 수 있는 정보량이 제한적이다. 또한 기존 복수영상 기반의 초고해상도 기법은 단일 영상을 사용했을 경우보다 영상의 보간 시 많은 정보를 사용할 수 있으나 영상의 내용에 따라 기법의 적용이 제한적이고, 컷(cut)의 경계 부근에서 기법의 성능이 매우 떨어지는 단점이 있다. 제안된 기법에서는 컷 검출(cut-detection) 기법을 통해 각 장면의 경계부근에서 적응적으로 단일영상 기반의 초고해상도 기법을 사용한다. 또한 움직임 벡터의 정규화 및 블록 단위의 윤곽선(edge) 패턴 분석을 통해 여러 제한조건에 강한 복수 영상 기반의 초고해상도 기법을 제안한다. 실험을 통하여 제안된 기법이 객관적, 주관적으로 기존의 기법보다 우수한 성능을 보이는 것을 확인하였다.