• Journal of Multimedia Information System
• /
• 제1권1호
• /
• pp.87-93
• /
• 2014

The aim of microscopic image restoration is to recover the image by applying the inverse process of degradation, and the results facilitate automated and improved analysis of the image. In this work, we consider the problem of image restoration as a minimization problem of convex cost function, which consists of a least-squares fitting term and regularization terms with non-negative constraints. The finite step method is proposed to solve this constrained convex optimization problem. We demonstrate the convergence of this method. Efficiency and restoration capability of the proposed method were tested and illustrated through numerical experiments.

• 한국컴퓨터정보학회논문지
• /
• 제13권2호
• /
• pp.121-125
• /
• 2008

멀티미디어와 관련한 소프트웨어 및 하드웨어의 기술적 발달로 인해 자료로부터 추출된 정보를 저장하거나 표현하는 수단으로써 영상이 사용되어지고 있다. 특히 유비쿼터스 환경에서는 이들 영상을 형성하고, 감지하며, 기록 및 전송하는 처리과정에서 다양한 원인으로 인해 노이즈(noise)가 부가된다. 이들 노이즈를 제거하기 위해 영상 복원에서는 각 노이즈 특성에 맞도록 적합한 필터링을 활용하게 된다. 뿐만 아니라 저해상도의 여러 영상들을 고해상도 영상으로 복원할 수 있는 방법 및 노이즈 혹은 훼손된 영상을 복원하는 다양한 방법과 기법들이 연구되고 있다. 이에 본 연구에서는 유비쿼터스 환경 및 의료 분야에 빈번히 발생될 수 있는 훼손된 영상으로부터 최적의 영상복원 매개변수를 적용하여 영상을 개선시키기 위한 방법으로, 사용자가 정의한 평균 필터와 위너 필터를 이용한 영상 개선 프로토타입 시스템을 구축하며, 구축된 시스템으로부터 훼손된 영상을 개선된 영상과 비교하여 노이즈의 특성과 개선된 결과를 보인다.

• 한국언론정보학보
• /
• 제38권
• /
• pp.73-118
• /
• 2007

이 연구의 목적은 위기 상황에서 조직이 내세우는 이미지 회복 전략이 대부분 언론을 통해 전달된다는 점에 착안하여, 언론보도 프레임이 공중의 이미지 회복 전략에 대한 판단에 영향을 주는가를 살펴보는 것이다. 이러한 연구 목적을 달성하기 위해서 신문을 중심으로 한 내용분석과 일반인을 대상으로 한 실험연구를 진행했다. 특히, 이 연구에서는 구체적인 위기 상황을 중심으로 조직의 이미지 회복 전략과 언론보도 프레임의 영향력을 알아보고자 하였으며, 위기 사건 중에서도 산업 전반 및 공중의 삶에 큰 영향을 주는 대표적인 사회 '위기'인 식품관련 사건을 선정하였다. 신문 내용분석 결과, 조직의 이미지 회복 전략은 대부분 일화중심 기사를 통해 공중에게 전달되는 경향을 보이고 있었다. 또한, 조직들이 구사하는 대부분의 이미지 회복 전략은 방어적인 것으로 이는 위기의 유형에 따른 이미지 회복 전략의 차별성을 보여주지 못하는 결과이다. 실험 결과, 조직들의 이미지 회복 전략은 공중이 선호하는 전략과 큰 차이를 보였으며 이러한 공중들의 인식은 프레임 유형에 따라 변화하는 양상을 보였다.

• Journal of information and communication convergence engineering
• /
• 제7권2호
• /
• pp.231-234
• /
• 2009

Speckle noise reduction for power Doppler ventricle coherent image for restoration and enhancement using Fast Wavelet Transform with multi-thresholding and multi-filtering on the each subbands is presented. Fast Wavelet Transform divides into low frequency component image to high frequency component image to be multi-resolved. Speckle noise is located on high frequency component in multi-resolution image mainly. A Doppler ventricle image is transformed and inversed with separated threshold function and filtering from low to high resolved images for restoration to utilize visualization for ventricle diagnosis. The experimental result shows that the proposed method has better performance in comparison with the conventional method.

• IEIE Transactions on Smart Processing and Computing
• /
• 제3권2호
• /
• pp.41-51
• /
• 2014

This paper describes a method to estimate the noise power using the minimum statistics approach, which was originally proposed for audio processing. The proposed minimum statistics-based method separates a noisy image into multiple frequency bands using the three-level discrete wavelet transform. By assuming that the output of the high-pass filter contains both signal detail and noise, the proposed algorithm extracts the region of pure noise from the high frequency band using an appropriate threshold. The region of pure noise, which is free from the signal detail part and the DC component, is well suited for minimum statistics condition, where the noise power can be extracted easily. The proposed algorithm reduces the computational load significantly through the use of a simple processing architecture without iteration with an estimation accuracy greater than 90% for strong noise at 0 to 40dB SNR of the input image. Furthermore, the well restored image can be obtained using the estimated noise power information in parametric image restoration algorithms, such as the classical parametric Wiener or ForWaRD image restoration filters. The experimental results show that the proposed algorithm can estimate the noise power accurately, and is particularly suitable for fast, low-cost image restoration or enhancement applications.

• 한국정보통신학회:학술대회논문집
• /
• 한국해양정보통신학회 2009년도 춘계학술대회
• /
• pp.156-159
• /
• 2009

Speckle noise reduction for 3D power doppler ventricle coherent image for restoration and enhancement using wavelet packet transform with separated thresholding is presented. Wavelet Packet Transform divide into low frequency component image to high frequency component image to be multi-resolved. speckle noise is located on high frequency component in multiresolution image mainly. A ventricle image is transformed and inversed with separated threshold function from low to high resolved images for restoration to be utilize visualization for ventricle diagnosis. The experimental result shows that the proposed method has better performance in comparison with the conventional method.

• Communications for Statistical Applications and Methods
• /
• 제29권1호
• /
• pp.27-40
• /
• 2022

In this paper, the focus on the removal noise in the binary image based on the variational Bayesian method with the Ising model. The observation and the latent variable are the degraded image and the original image, respectively. The posterior distribution is built using the Markov random field and the Ising model. Estimating the posterior distribution is the same as reconstructing a degraded image. MCMC and variational Bayesian inference are two methods for estimating the posterior distribution. However, for the sake of computing efficiency, we adapt the variational technique. When the image is restored, the iterative method is used to solve the recursive problem. Since there are three model parameters in this paper, restoration is implemented using the VECM algorithm to find appropriate parameters in the current state. Finally, the restoration results are shown which have maximum peak signal-to-noise ratio (PSNR) and evidence lower bound (ELBO).

• 문화기술의 융합
• /
• 제4권3호
• /
• pp.241-246
• /
• 2018

영상 기기의 기계적인 문제로 인해 실험 데이터에 발생한 잡음으로 인한 인식 오류를 최소화하기 위해서 영상복원 과정을 거친다. 영상 복원 방법은 실험 데이터를 구성하는 각각의 픽셀에 대해 Direct Neighbor와 Indirect Neighbor의 개수와 위치를 조사해서 잡음을 해결한다. 결과적으로, 영상 복원 과정을 통해 실험 데이터에 발생한 잡음을 최대한 제거하고, 영역 단위로 학습 데이터와 실험 데이터의 차이를 계산해서 잡음에 의한 인식 오류 가능성을 낮춤으로써 만족할만한 인식 결과를 얻을 수 있다.

• ETRI Journal
• /
• 제41권4호
• /
• pp.415-425
• /
• 2019

Generative adversarial networks (GAN) have been successfully used in many image restoration tasks, including image denoising, super-resolution, and compression artifact reduction. By fully exploiting its characteristics, state-of-the-art image restoration techniques can be used to generate images with photorealistic details. However, there are many applications that require faithful rather than visually appealing image reconstruction, such as medical imaging, surveillance, and video coding. We found that previous GAN-training methods that used a loss function in the form of a weighted sum of fidelity and adversarial loss fails to reduce fidelity loss. This results in non-negligible degradation of the objective image quality, including peak signal-to-noise ratio. Our approach is to alternate between fidelity and adversarial loss in a way that the minimization of adversarial loss does not deteriorate the fidelity. Experimental results on compression-artifact reduction and super-resolution tasks show that the proposed method can perform faithful and photorealistic image restoration.

• 융합신호처리학회논문지
• /
• 제13권4호
• /
• pp.187-193
• /
• 2012

An image authentication scheme for gray scale image through embedding a digital watermark by quantization of Integer Wavelet Transform (IWT) coefficients of the image is proposed in this paper. Proposed method is designed to detect modification of an image and to identify tampered location of the image. To embed the watermark mid-frequency band of a second level IWT was used. An approximation of the original image based on LL band was stored in LSB bits of the pixel data as a recovery mark for restoration of the image. Watermarked image has achieved a good PSNR of 40 dB compared to original cover image. Restored image quality was also very good with a PSNR of more than 35 dB compared to unmodified watermarked image even when 25% of the received image is cropped. Thus, the proposed method ensures a proper balance between the fidelity of the watermarked image and the quality of the restored image.