• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2011.11a
• /
• pp.129-130
• /
• 2011

본 논문에서는 입체감을 보존하여 스테레오 영상의 크기를 조정하는 기법을 제안한다. 스테레오 영상의 디스패리티 데이터를 이용하여 좌우 영상의 중요 영역을 추출하고, 영상의 빈 공간을 인페인팅(inpainting) 기법을 통해 복원한다. 복원된 배경 영역의 크기를 선형적으로 조정하고, 중요 영역은 최대한 보존하도록 크기를 조정하여 배경 영역에 붙여 넣는다. 실험 결과 제안하는 알고리즘이 기존 이미지 크기 조정 기법보다 스테레오 영상의 중요 정보와 입체감을 왜곡하지 않고 보존함을 확인할 수 있었다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2019.06a
• /
• pp.5-8
• /
• 2019

단일 영상 초해상도 (Single Image Super-Resolution - SISR)기법은 카메라로 획득된 저해상도 영상에 필터 기반의 연산을 적용하여 좋은 화질의 고해상도 영상을 복원하는 과정이다. 최근에 심층 합성곱 신경망 학습의 발전에 따라 단일 영상 초해상도에 적용되는 심층 학습 기법들은 좋은 성과를 보여 주고 있다. 본 논문은 단일 영상 초해상도 성능을 개선하기 위해 웨이블릿 예측 네트워크를 효율적으로 적용하는 방법에 대해 연구하였으며, 저해상도 입력 영상의 특징을 잘 추출해내기 위해 네트워크 내부에 RDB를 적용하여 기존 방식보다 효율적으로 고해상도 영상 복원하는 기법을 제안한다. 모의실험을 통해 제안하는 방법이 기존 방법보다 화질은 약 PSNR 0.18dB만큼 우수하며 속도는 1.17배 빠른 것을 확인하였다.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.49 no.11
• /
• pp.47-55
• /
• 2012

In this paper, we propose a method to reconstruct accurate 3D shape object by using multi-view images which are disturbed by motion blur. In multi-view deblurring, more precise PSF estimation can be done by using the geometric relationship between multi-view images. The proposed method first estimates initial 2D PSFs from individual input images. Then 3D PSF candidates are projected on the input images one by one to find the best one which are mostly consistent with the initial 2D PSFs. 3D PSF consists with direction and density and it represents the 3D trajectory of object motion. 야to restore 3D shape by using multi-view images computes the similarity map and estimates the position of 3D point. The estimated 3D PSF is again projected to input images and they replaces the intial 2D PSFs which are finally used in image deblurring. Experimental result shows that the quality of image deblurring and 3D reconstruction improves significantly compared with the result when the input images are independently deblurred.

• KIPS Transactions on Software and Data Engineering
• /
• v.9 no.12
• /
• pp.403-410
• /
• 2020

In addition to the research on noise removal and super-resolution using the data restoration (Output result) function of Autoencoder, research on the performance improvement of clustering using the dimension reduction function of autoencoder are actively being conducted. The clustering function and data restoration function using Autoencoder have common points that both improve performance through the same learning. Based on these characteristics, this study conducted an experiment to see if the autoencoder model designed to have excellent data recovery performance is superior in clustering performance. Skip connection technique was used to design autoencoder with excellent data recovery performance. The output result performance and clustering performance of both autoencoder model with Skip connection and model without Skip connection were shown as graph and visual extract. The output result performance was increased, but the clustering performance was decreased. This result indicates that the neural network models such as autoencoders are not sure that each layer has learned the characteristics of the data well if the output result is good. Lastly, the performance degradation of clustering was compensated by using both latent code and skip connection. This study is a prior study to solve the Hanja Unicode problem by clustering.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.19 no.3
• /
• pp.119-126
• /
• 2005

Electrical resistance tomography(ERT) maps resistivity values of the soil subsurface and characterizes buried objects. The characterization includes location, size and resistivity of buried objects. In this paper, Gauss-Newton, truncated least squares(TLS) and simultaneous iterative reconstruction technique(SIRT) methods are presented for the solution of the ERT image reconstruction. Computer simulations show that the spatial resolution of the reconstructed images by the TLS approach is improved as compared to those obtained by the Gauss-Newton and SIRT method.

• Journal of the Korean Geographical Society
• /
• v.33 no.3
• /
• pp.435-446
• /
• 1998

This Paper reports that a new remote sensing techlique focused on a narrow and long strip target (e.g. 15m wide and 100km long) has been specifically developed for pipeline ROW (Right-Of-Way) recovery monitoring. With video it was possible to isolate the maior vegetation communities of the narrow pipeline ROW with acceptable spatial precision by visual or quantitative methods. It was particuarly useful when used to assess a variety of spatial patch dynamics for ROW recovery through digital change-detection techniques in a GIS environment. The main conclusion of this paper is that VSM is a realistic operational technique for a pipeline ROW application. The results also indicate that VSM could be extensively used for other examples of linear thematic mapping.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2004.05b
• /
• pp.1340-1344
• /
• 2004

우리나라는 계절에 따라 수량의 변화가 커서 하천에 흐르는 유량의 변화가 심하지만 대부분의 하천은 갈수기에도 각각 일정량의 유량을 가지고 흘러간다. 그러나 충적층을 통과하는 소하천의 경우에는 계절에 상관없이 하천수가 지하로 침투하여 지표수가 감소한다. 소하천의 하천수량이 갈소되거나 건천화 될 경우 수질이 악화되거나 하천의 바닥이 드러날 경우 수중생태계는 파괴됨 것이다. 현재 지하수를 개발하여 부족한 하천수를 보충하는 방법을 모색하고 있으나 하천수의 지하침투가 심각한 경우에는 지하수만으로는 하천수의 부족분을 보충하는데 한계가 있다. 충적층을 통과하는 소하천의 침투수를 차단하는 방법으로 충적층에 불투수 매트를 설치하여 사용하면 하천의 건천화를 방지할 수 있으며, 충적층의 침투수를 차단하기 위하여 충적층위에 점토층을 설치하여 투수계수를 저하시키는 방법을 사용하면 부족한 하천유지용수를 확보할 수 있다. 따라서 충적층을 통과하는 소하천의 건천화를 방지하고 자연형 하천으로 개발할 경우 수중생태계 복원성이 우수하며, 자연경관의 복원에도 효과가 있다.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.47 no.5
• /
• pp.145-152
• /
• 2010

This paper presents a digital auto-focusing algorithm using iterative unsharp mask filter. The proposed digital auto-focusing algorithm has the advantage of low computational complexity because it uses a simple filter instead of calculating the point spread function for the estimation of image degradation. The proposed iterative algorithm can control the number of iterations for image restoration according to the objective and the subjective criterion. We show that the proposed algorithm is mathematically equivalent to the conventional image restoration. Finally, in order to evaluate the performance of the proposed algorithm, various experiments are performed so that the proposed algorithm can provide good results in the sense of subjective and objective views.

• The KIPS Transactions:PartA
• /
• v.13A no.7 s.104
• /
• pp.639-646
• /
• 2006

Data compression technologies enable us to store and transfer large amount of data efficiently, and become more and more important due to increasing data size and the network traffic. Moreover, as a result of the increase of computing power, volumetric data produced from various applied science and engineering fields has been getting much larger. In this Paper, we present a volume compression scheme which exploits Daubeches wavelet transform. The proposed scheme basically supports lossy compression for 3D volume data, and provides unit-wise random accessibility. Since our scheme shows far lower error rates than the previous compression methods based on Haar filter, it could be used well for interactive visualization applications as well as large volume data compression requiring image fidelity.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.11 no.6
• /
• pp.1122-1128
• /
• 2007

The denoising for image restoration based on the Wavelet Packet Transform with AA(Absolute Average) making-threshold is presented. The wavelet packet transform leads to be better in the part of high frequency than wavelet transform to eliminate noise. And the existing threshold determination is used standard deviation estimated results in increasing the noise and threshold, and damaging an image quality. In addition that is decreased image restoration PSNR by using the same threshold in spite of changing image because of installing a threshold in proportion of noise size. In contrast the AA thresholding method with wavelet packet is adapted by changing image to set up threshold by statistic quantity of resolved image and is avoided an extreme impact. The results on the experiment has improved 10% and 5% over than the denoising based on simple wavelet transform and wavelet packet respectively.