• ETRI Journal
• /
• 제32권3호
• /
• pp.390-394
• /
• 2010

In this paper, we propose a new super-resolution technique based on interpolation of the high-frequency subband images obtained by discrete wavelet transform (DWT) and the input image. The proposed technique uses DWT to decompose an image into different subband images. Then the high-frequency subband images and the input low-resolution image have been interpolated, followed by combining all these images to generate a new super-resolved image by using inverse DWT. The proposed technique has been tested on Lena, Elaine, Pepper, and Baboon. The quantitative peak signal-to-noise ratio (PSNR) and visual results show the superiority of the proposed technique over the conventional and state-of-art image resolution enhancement techniques. For Lena's image, the PSNR is 7.93 dB higher than the bicubic interpolation.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 2002년도 춘계학술대회 논문집
• /
• pp.203-206
• /
• 2002

One of the ever-increasing demands on the performances of heterodyne interferometers is to improve the measurement resolution, of which current state-of-the-art reaches the region of sub-nanometers. We propose a new scheme of phase-measuring electronics that reduces the measurement resolution without further increase in clock speed. Our scheme adopts a super-heterodyne technique that lowers the original beat frequency to a level of 1 MHz by mixing it with electrically generated reference signal. The technique enables us to measure the phase of Doppler shift with a resolution of 1.58 nanometer at a sampling rate of 1 MHz. To avoid the undesirable decrease in the maximum measurable speed caused by the lowered beat frequency, a special from of frequency up-down counting technique is combined with the super-heterodyning. This alloys performing required phase unwrapping simply by using programmable digital gates without 2$\pi$ ambiguities up to the maximum velocity of 2.35 m/s.

• 스마트미디어저널
• /
• 제8권4호
• /
• pp.64-71
• /
• 2019

The performance of an OMR (Optical Music Recognition) system is usually determined by the characterizing features of the input music score images. Low resolution is one of the main factors leading to degraded image quality. In this paper, we handle the low-resolution problem using the super-resolution technique. We propose the use of a deep neural network with instance normalization to improve the quality of music score images. We apply instance normalization which has proven to be beneficial in single image enhancement. It works better than batch normalization, which shows the effectiveness of shifting the mean and variance of deep features at the instance level. The proposed method provides an end-to-end mapping technique between the high and low-resolution images respectively. New images are then created, in which the resolution is four times higher than the resolution of the original images. Our model has been evaluated with the dataset "DeepScores" and shows that it outperforms other existing methods.

• 한국자동차공학회논문집
• /
• 제22권7호
• /
• pp.31-39
• /
• 2014

Recently, as Intelligent Transportation System (ITS) and self-driving system become influential in the ground transportation system, automotive radar systems have been actively studied among the various radar systems to implement the vehicle collision detection system and distance measurement system between vehicles. Most of the automotive radars are Frequency Modulated Continuous Wave (FMCW) radar type which can calculate distance and velocity of target by estimating the frequency difference between the transmitted signal and received signal. Therefore, accurate frequency estimation is very important in the FMCW radar system. For this reason, to improve the measurement accuracy of the FMCW radar, Reverse Directional FRI (RD-FRI) Super-Resolution technique which has high frequency estimation accuracy is applied to the FMCW radar system. The feasibility of the proposed technique is evaluated with simulation results and compared with FFT and conventional Super-Resolution techniques. The simulation results show that the proposed technique estimates the frequency with high accuracy and the distance with centimeter accuracy.

• 한국정보통신학회논문지
• /
• 제25권6호
• /
• pp.774-784
• /
• 2021

객체 검출 및 인식 과정은 컴퓨터비전 분야에서 매우 중요한 과업으로써, 관련 연구가 활발하게 진행되고 있다. 그러나 실제 객체 인식 과정에서는 학습된 이미지 데이터와 테스트 이미지 데이터간 해상도 차이로 인하여 인식기의 정확도 성능이 저하되는 문제가 종종 발생한다. 이를 해결하기 위해 본 논문에서는 객체 인식 정확도 향상을 위한 이미지 초해상도 기법을 제안하여 객체 인식 및 초해상도 통합 프레임워크를 설계하고 개발하였다. 세부적으로는 11,231장의 차량 번호판 훈련용 이미지를 웹 크롤링, 인조데이터 생성 등을 통해 자체적으로 구축하고, 이를 활용하여 이미지 좌우 반전에 강인하도록 목적함수를 정의하여 이미지 초해상도 인공 신경망을 훈련시켰다. 제안 방법의 성능을 검증하기 위해 훈련된 이미지 초해상도 및 번호 인식기 1,999장의 테스트 이미지에 실험하였고, 이를 통해 제안한 초해상도 기법이 문자 인식 정확도 개선 효과가 있음을 확인하였다.

• 한국컴퓨터그래픽스학회논문지
• /
• 제25권3호
• /
• pp.105-113
• /
• 2019

물리 기반 유체 시뮬레이션은 고해상도 연산을 위해 많은 시간이 필요하다. 이 문제를 해결하기 위해 저해상도 유체 시뮬레이션의 한계를 딥 러닝으로 보완하는 연구들이 있으며, 그중에서는 저해상도의 시뮬레이션 데이터를 고해상도로 변환해주는 Super-resolution 분야가 있다. 하지만 기존 기법들은 전체 데이터 공간에서 밀도 데이터가 없는 부분까지 연산하므로 전체 시뮬레이션 속도 면에서 효율성이 떨어지며, 입력 해상도가 큰 경우에는 GPU 메모리가 부족해 연산할 수 없는 경우가 발생할 수 있다. 본 연구에서는 공간 분할 법 중 하나인 쿼드 트리를 활용하여 시뮬레이션 공간을 분할 및 분류하여 Super-resolution 하는 기법을 제안한다. 본 기법은 필요 공간만 Super-resolution 하므로 전체 시뮬레이션 가속화가 가능하고, 입력 데이터를 분할 연산하므로 GPU 메모리 문제를 해결할 수 있게 된다.

• 한국수자원학회논문집
• /
• 제56권10호
• /
• pp.641-653
• /
• 2023

기후변화와 도시화의 영향으로 인해 자연재해의 발생빈도와 규모가 증가하고 있다. 특히 도시 침수는 발생 시간이 짧고 막대한 인명 및 경제적 손실을 초래할 수 있기 때문에 신속하고 정확도 높은 예측 정보 생산이 중요하다. 하지만, 기존 물리과정 및 인공지능 기반 기법은 고해상도 침수 해석을 위해 많은 전산 자원이나 데이터가 요구되는 한계가 있다. 본 연구에서는 딥러닝 기반 초해상화(Super-Resolution) 기법을 통한 고해상도 도시 침수 해석 방법을 제안하고 적용성을 평가한다. 제안된 방법은 고해상도 물리 모형의 결과로 훈련된 초해상화 딥러닝 모형을 이용하여 저해상도 침수 해석 이미지를 고해상도로 변환한다. 미국 포틀랜드 도심지의 두 가지 침수 사례에 대해 적용, 4 m 공간해상도 물리 모의 결과를 1 m 급 고해상도 침수 해석 정보로 초해상화 하였으며, 초해상화 이미지와 고해상도 원본 간 높은 구조적 유사성이 확인되었다. 성능 지표로 평가한 결과, 전체 검증 대상 이미지에 대한 평균 PSNR 22.77 dB, SSIM 0.77로 우수하여, 초해상화 기법의 도시 침수 해석 적용성이 검증되었다. 제안된 방법은 적은 양의 침수 시나리오만으로도 효율적인 딥러닝 모형 훈련이 가능하고, 물리 모형의 정보를 최대한 활용할 수 있기 때문에, 고해상도 도시 침수 정보 생산에 효과적으로 사용될 수 있을 것으로 기대된다.

• 한국정밀공학회지
• /
• 제19권9호
• /
• pp.172-178
• /
• 2002

One of the ever-increasing demands on the performances of heterodyne interferometers is to improve the measurement resolution, of which current state -of-the-art reaches the region of sub-nanometers. So far, the demand has been met by increasing the clock speed that drives the electronics involved fur the phase measurement of the Doppler shift, but its further advance is being hampered by the technological limit of modem electronics. To cope with the problem, in this investigation, we propose a new scheme of phase -measuring electronics that reduces the measurement resolution without further increase in clock speed. Our scheme adopts a super-heterodyne technique that lowers the original beat frequency to a level of 1 MHz by mixing it with a stable reference signal generated from a special phase- locked-loop. The technique enables us to measure the phase of Doppler shift with a resolution of 1.58 nanometer at a sampling rate of 1 MHz. To avoid the undesirable decrease in the maximum measurable speed caused by the lowered beat frequency, a special form of frequency up-down counting technique is combined with the super-heterodyning. This allows performing required phase unwrapping simply by using programmable digital gates without 2n ambiguities up to the maximum velocity guaranteed by the original beat frequency.

• 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.

• 한국측량학회지
• /
• 제31권6_2호
• /
• pp.559-565
• /
• 2013

Spatial resolution of land covers from remotely sensed images can be increased using super-resolution mapping techniques for soft-classified land cover proportions. A further development of super-resolution mapping technique is downscaling the original remotely sensed image using super-resolution mapping techniques with a forward model. In this paper, the model for increasing spatial resolution of remote sensing multispectral image is tested with real SPOT 5 imagery at 10m spatial resolution for an area in Bac Giang Province, Vietnam in order to evaluate the feasibility of application of this model to the real imagery. The soft-classified land cover proportions obtained using a fuzzy c-means classification are then used as input data for a Hopfield neural network (HNN) to predict the multispectral images at sub-pixel spatial resolution. The 10m SPOT multispectral image was improved to 5m, 3,3m and 2.5m and compared with SPOT Panchromatic image at 2.5m resolution for assessment.Visually, the resulted image is compared with a SPOT 5 panchromatic image acquired at the same time with the multispectral data. The predicted image is apparently sharper than the original coarse spatial resolution image.