• 한국멀티미디어학회논문지
• /
• 제18권6호
• /
• pp.691-700
• /
• 2015

Here, we present a new framework for histogram equalization in which both local and global contrasts are enhanced using neighborhood metrics. When checking neighborhood information, filters can simultaneously improve image quality. Filters are chosen depending on image properties, such as noise removal and smoothing. Our experimental results confirmed that this does not increase the computational cost because the filtering process is done by our proposed arrangement of making the histogram while checking neighborhood metrics simultaneously. If the two methods, i.e., histogram equalization and filtering, are performed sequentially, the first method uses the original image data and next method uses the data altered by the first. With combined histogram equalization and filtering, the original data can be used for both methods. The proposed method is fully automated and any spatial neighborhood filter type and size can be used. Our experiments confirmed that the proposed method is more effective than other similar techniques reported previously.

• 한국기계가공학회지
• /
• 제16권1호
• /
• pp.1-8
• /
• 2017

Warpage is a major defect frequently found in the injection molding process, and the reduction of warpage is a very challenging problem because of the uncontrollable factors, such as variations in the process parameters. Without any countermeasure against these noises, attempts to reduce the defects often lead to failure. In this research, a new robust design methodology, based on the Mahalanobis Taguchi System (MTS) to reduce warpage, is presented. The MTS performs the orthogonal array experiments and uses the signal-to-noise (SN) ratio of the Mahalanobis distance as a performance metric. The validity of the proposed method is illustrated through an optimal design of the injection molding process of a CPU base plate.

• 전자공학회논문지S
• /
• 제35S권4호
• /
• pp.85-95
• /
• 1998

In this paper, a new watermarking method for a copy protectionof images is proposed. The proposed method adaptively embeds a watermark in the frequency domain of images using human visual system model. For this purpose, the Just Noticeable Differences(JNDs) of each frequency coeffeicient value of a luminance plane is first found using Watson and Solomon's visual system model. An invisible maximum watermark value with is different in every position according to the characteristics of images is determined usig JND and Minkowski metric. A low frequency domain is divided into two sets based on a PN-sequence to protect thewatermark from the attack. The watermarks are added to one set of coefficients and detecting a watermark, the difference between the mean values of absolute coefficient values of both sets is calculated. The embedded watermark is tested using statistical hypothesis based on test static dertermined by the ean difference. To demonstrate the perfromance of the proposed method, the new watermarking method is applied to a high frequency image and low frequency images. Experimenatal results show the watermark is invisible and robust to JPEGlossy compression and noise.

• 한국정보디스플레이학회:학술대회논문집
• /
• 한국정보디스플레이학회 2009년도 9th International Meeting on Information Display
• /
• pp.1224-1227
• /
• 2009

When evaluating the quality of images and displays, it is important to combine the characteristics as perceived by the human visual system and measured by equipment using subjective and objective methods, respectively. In the case of objective methods, the quality of a display is measured using colorimetric or radiometric devices according to existing standards covering the color temperature, gamut size, gamma characteristic, and device characterization. Meanwhile, subjective methods assess the quality of an image using the human visual system based on a comparison with a reference or counterpart using such metrics as the sharpness, noise, contrast, saturation, and color accuracy. Objective and subjective methods are usually used together in comparison, as ultimately it is observers watching images on a display. In addition to existing objective methods, a new image quality metric is also introduced as regards the JPEG compression ratio that is reflected in the relationship between the gamut size and the color fidelity in CIELAB color space.

• Journal of the Korean Society for Industrial and Applied Mathematics
• /
• 제25권2호
• /
• pp.26-38
• /
• 2021

Image coloration refers to adding plausible colors to a grayscale image or video. Image coloration has been used in many modern fields, including restoring old photographs, as well as reducing the time spent painting cartoons. In this paper, a method is proposed for colorizing grayscale images using a convolutional neural network. We propose an encoder-decoder model, adapting FusionNet to our purpose. A proper loss function is defined instead of the MSE loss function to suit the purpose of coloring. The proposed model was verified using the ImageNet dataset. We quantitatively compared several colorization models with ours, using the peak signal-to-noise ratio (PSNR) metric. In addition, to qualitatively evaluate the results, our model was applied to images in the test dataset and compared to images applied to various other models. Finally, we applied our model to a selection of old black and white photographs.

• 대한수학회지
• /
• 제60권2호
• /
• pp.255-271
• /
• 2023

In this paper, we deal with random attractors for dynamical systems forced by a deterministic noise. These kind of systems are modeled as skew products where the dynamics of the forcing process are described by the base transformation. Here, we consider skew products over the Bernoulli shift with the unit interval fiber. We study the geometric structure of maximal attractors, the orbit stability and stability of mixing of these skew products under random perturbations of the fiber maps. We show that there exists an open set U in the space of such skew products so that any skew product belonging to this set admits an attractor which is either a continuous invariant graph or a bony graph attractor. These skew products have negative fiber Lyapunov exponents and their fiber maps are non-uniformly contracting, hence the non-uniform contraction rates are measured by Lyapnnov exponents. Furthermore, each skew product of U admits an invariant ergodic measure whose support is contained in that attractor. Additionally, we show that the invariant measure for the perturbed system is continuous in the Hutchinson metric.

• 전자공학회논문지CI
• /
• 제45권3호
• /
• pp.149-159
• /
• 2008

참조 영상을 이용하여 영상의 분위기를 전환하고자 할 때, 영상의 분위기에 영향을 주는 요소 중 하나인 색을 이용하여 영상이 가지는 분위기를 변환한다. 색을 변환할 때, 국부적인 색상의 특징을 반영하기 위해 입력 영상의 화소에 대하여 색인을 하는데 낮은 채도 상에서는 색상의 식별력 저하로 인해 부적절한 색인의 색상 항목으로 색인이 되는 문제가 발생한다. 따라서 본 논문에서는 유채색과 무채색을 분리하여 처리함으로써, 낮은 채도 상에서의 잘못된 색인이 일어나지 않도록 제한하는 기법을 제안한다. 제안하는 기법은 채도의 문턱치를 이용하여 유채색과 무채색을 구분하는 단계, 구분된 화소들을 실린드리컬 거리(Cylindrical metric)를 이용하여 11가지 색상 항목(Index color)으로 색인을 하는 단계, 각 색상 항목의 우선순위 결정하고 평균과 표준편차를 구하는 단계, 마지막으로 Lab 색 공간에서 색을 변환하고 영상의 잡음과 의사 윤곽선(pseudo-contour)을 제거하기 위한 후처리 단계의 4단계로 구성된다. 실험결과를 통해 제안하는 기법은 낮은 채도 상에서도 유채색과 무채색이 잘 분리되어 색인이 되었으며 원본 영상의 색이 참조 영상의 색으로 자연스럽게 변환된 결과를 얻을 수 있음을 확인하였다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제17권3호
• /
• pp.881-895
• /
• 2023

Modern image inpainting techniques based on deep learning have achieved remarkable performance, and more and more people are working on repairing more complex and larger missing areas, although this is still challenging, especially for facial image inpainting. For a face image with a huge missing area, there are very few valid pixels available; however, people have an ability to imagine the complete picture in their mind according to their subjective will. It is important to simulate this capability while maintaining the identity features of the face as much as possible. To achieve this goal, we propose a three-stage network model, which we refer to as the identity and structure feature refinement network (ISFRNet). ISFRNet is based on 1) a pre-trained pSp-styleGAN model that generates an extremely realistic face image with rich structural features; 2) a shallow structured network with a small receptive field; and 3) a modified U-net with two encoders and a decoder, which has a large receptive field. We choose structural similarity index (SSIM), peak signal-to-noise ratio (PSNR), L1 Loss and learned perceptual image patch similarity (LPIPS) to evaluate our model. When the missing region is 20%-40%, the above four metric scores of our model are 28.12, 0.942, 0.015 and 0.090, respectively. When the lost area is between 40% and 60%, the metric scores are 23.31, 0.840, 0.053 and 0.177, respectively. Our inpainting network not only guarantees excellent face identity feature recovery but also exhibits state-of-the-art performance compared to other multi-stage refinement models.

• 한국음향학회지
• /
• 제37권3호
• /
• pp.147-155
• /
• 2018

본 연구의 목적은 선형 다중회귀분석을 통해 과부하 굴삭기 소음의 불쾌감을 표현할 수 있는 음질인자를 찾고 이를 이용하여 객관적인 평가를 위한 음질인덱스를 개발하고자 하였다. 이를 위해 소음특성이 다른 굴삭기 6모델을 선별하여 과부하 조건의 운전실 소음을 녹음하고 쌍대비교법과 크기평가법에 의한 청음평가를 수행하였다. 그리고 일관성이 낮은 청음평가 결과는 군집분석을 통해 선호도 성향이 다른 2 그룹으로 분류하고 회귀분석을 실시하여 어떤 음질인자가 불쾌감(낮은 선호도)에 유의한 영향을 주는 인자인지를 고찰하였다. 그 결과 불쾌감을 표현하는 음질인자는 10 Bark까지 부분 라우드니스 인자($PN_{10Bark}$)에 따라 불쾌감을 느끼는 성향을 가진 청음 평가자 그룹과 엔진 소음($dB_{EG}$)과 유압시스템 소음의 차이($dB_1$)로 표현된 인자에 따라 불쾌감을 느끼는 성향을 가진 청음 평가자 그룹으로 구성되어 있음을 확인하였다. 그리고 쌍대비교법의 선호도 순위와 성향 분석결과를 이용하고 크기평가법과의 상관분석을 통해 신뢰도가 낮은 평가자의 결과는 제외하여 보다 신뢰성 높은 크기평가법의 청음평가 결과를 얻었다.

• 한국정보과학회논문지:소프트웨어및응용
• /
• 제37권1호
• /
• pp.1-8
• /
• 2010

최근 마이크로어레이 데이터를 기반으로 두 개의 샘플 그룹간에 유의한 발현 차이를 나타내는 생물학적 기능 그룹을 검출하기 위한 유전자 집합 분석(gene set analysis) 연구가 많은 주목을 받고 있다. 기존의 유의 유전자 검출 연구와는 달리, 유전자 집합 분석 연구는 유의한 유전자 집합과 이들의 기능적 특징을 함께 검출할 수 있다는 장점이 있다. 이러한 이유로 최근에는 PAGE, GSEA 등과 같은 다양한 통계적 방식의 유전자 집합 분석 방법들이 소개되고 있다. 특히, PAGE의 경우 두 샘플 그룹간의 유전자 발현 차이를 나타내는 스코어의 분포가 정규 분포임을 가정하는 모수적 접근 방식을 취하고 있다. 이러한 방법은 GSEA 등과 같은 비모수적 방식에 비해 계산량이 적고 성능이 비교적 우수한 장점이 있다. 하지만, PAGE에서 유전자 발현 차이를 정량화하기 위한 메트릭으로 사용하고 있는 AD(average difference)의 경우, 두 그룹간에 절대적 평균 발현 차이만을 고려하기 때문에 실제 유전자의 발현값 크기나 분산의 크기에 따른 상대적 중요성을 반영하지 못하는 문제가 있다. 본 논문에서는 이를 보완하기 위해 실제 유전자의 발현값 크기나 그룹 내 샘플들의 분산 정보 등을 스코어 계산에 함께 반영하는 WAD(weighted average difference), FC(Fisher's criterion), 그리고 Abs_SNR(Absolute value of signal-to-noise ratio)을 모수적 방식의 유전자 집합 분석에 적용하고 이에 따른 유의 유전자 집합 검출 결과를 실험을 통해 비교 분석하였다.