• The Journal of the Korea Contents Association
• /
• 제9권5호
• /
• pp.91-98
• /
• 2009

Recent digital camera development tendency shows the higher ISO extension which enables a user to take a picture in deemed lighting condition without using a flash or another equipment. Also noise reduction function has been developed to adjust the higher ISO extension. But using the high ISO causes a lot of noises so that proper noise reduction function should be adjusted. So the manufacturers introduce the advanced noise reduction function with higher ISOs as the digital camera performance development. On the other hands, the defect of noise reduction is not widely open such as remaining noise in high ISOs and decreasing resolution. Therefore, in this research we verify and experiment the decrease of resolution by using noise reduction function on a digital camera. Moreover we analyze the results how much the noise reduction function effects to the image resolution. And we present the results as practical numerical values.

• Proceedings of the KIEE Conference
• /
• 대한전기학회 2003년도 춘계학술대회 논문집 전기기기 및 에너지변환시스템부문
• /
• pp.273-275
• /
• 2003

SMPS는 신속한 스위칭에 의하여 높은 $d{\upsilon}$/dt 및 di/dt 특성을 가지며, 이러한 스위칭 특성으로 SMPS에서 발생하는 EMI 노이즈는 기기 내부 및 외부에 간섭을 일으킨다. 일반적으로 전자파 이미션은 전도성 이미션 및 방사성 이 미션으로 분류되며. 전도성 이미션은 공통모드와 차동모드로 구분한다. 최근 노이즈전류를 평형시킴으로써 공통모드 노이즈를 감소시키는 방법이 연구되었다. 이 방법은 공통모드 노이즈가 시스템 불평형의 결과라는 측면을 고려한 것이며. 노이즈 전류의 평형상태에서 컨버터를 동작시키면 공통모드 노이즈가 크게 줄어들 수 있음을 보여준다. 본 논문에서는 절연형 SMPS에서 노이즈 전류의 평형기법을 이용하여 SMPS의 1차 및 2차 측의 노드 확장에 따른 노이즈 감소를 분석한다.

• Journal of the Korean Society of Radiology
• /
• 제18권2호
• /
• pp.93-99
• /
• 2024

The digital medical imaging, especially, computed tomography (CT), should necessarily be considered in terms of noise distribution caused by converting to X-ray photon to digital imaging signal. Recently, the denoising technique based on deep learning architecture is increasingly used in the medical imaging field. Here, we evaluated noise reduction effect according to various noise types based on the U-net deep learning model in the lung CT images. The input data for deep learning was generated by applying Gaussian noise, Poisson noise, salt and pepper noise and speckle noise from the ground truth (GT) image. In particular, two types of Gaussian noise input data were applied with standard deviation values of 30 and 50. There are applied hyper-parameters, which were Adam as optimizer function, 100 as epochs, and 0.0001 as learning rate, respectively. To analyze the quantitative values, the mean square error (MSE), the peak signal to noise ratio (PSNR) and coefficient of variation (COV) were calculated. According to the results, it was confirmed that the U-net model was effective for noise reduction all of the set conditions in this study. Especially, it showed the best performance in Gaussian noise.

• Proceedings of the Korean Magnestics Society Conference
• /
• 한국자기학회 2002년도 동계연구발표회 논문개요집
• /
• pp.182-183
• /
• 2002

고밀도 자기기록이 가능하기 위해서는 Ku가 크고 노이즈가 작은 매체가 필요하다. CoCrPt 기록층은 보자력이 크고 열적으로 안정하나 결정립크기가 크고 결정립간 교환 상호작용이 커서 노이즈가 크다.[1][2] 한편 CoCrPt 에 Ta, B 등 제4원소를 첨가하면 결정립이 크기가 감소하고 노이즈는 감소하나 Ku가 감소하여 열적으로 불안정해진다.[3] (중략)

• Journal of the Korean Society of Radiology
• /
• 제17권5호
• /
• pp.709-715
• /
• 2023

In this study, the performance evaluation of image quality for noise reduction was implemented using the U-net deep learning architecture in computed tomography (CT) images. In order to generate input data, the Gaussian noise was applied to ground truth (GT) data, and datasets were consisted of 8:1:1 ratio of train, validation, and test sets among 1300 CT images. The Adagrad, Adam, and AdamW were used as optimizer function, and 10, 50 and 100 times for number of epochs were applied. In addition, learning rates of 0.01, 0.001, and 0.0001 were applied using the U-net deep learning model to compare the output image quality. To analyze the quantitative values, the peak signal to noise ratio (PSNR) and coefficient of variation (COV) were calculated. Based on the results, deep learning model was useful for noise reduction. We suggested that optimized hyper parameters for noise reduction in CT images were AdamW optimizer function, 100 times number of epochs and 0.0001 learning rates.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• 제23권1호
• /
• pp.27-35
• /
• 2009

The moving picture blur noise of LCD is measured and the characteristics of it are evaluated. From the results, blur noise is generated only when the sight line moves while pursuing the moving image and blur noise is not generated when the sight line is fixed. In addition, decrease of gray level by the image blur has a linearity and velocity dependence. The blur noise simulator based on this experimental results is developed. From the results of blur noise simulation, the faster the moving speed of image is, the more blur noise has increased and these agree with the results of measurement. In the result of simulating blur noise characteristics by the duty ratio control of backlight, noise is reduced by lowering of the duty ratio. but the blur noise increases if there is in the both of adjacent two frames. Moreover, the case of doubling the frame rate to 120[Hz], decreasing the moving speed of the image by making an new image between the adjacent two frames brings the reduction of blur noise.

• Journal of the Korean Society of Radiology
• /
• 제13권5호
• /
• pp.765-771
• /
• 2019

Noise in skull X-ray imaging is inevitable, which reduces imaging quality and diagnostic accuracy and increases errors due to the nature of digital imaging devices. Increasing the dose can attenuate noise, but that could lead to big problems with higher exposure dose received by patients. Thus, noise reduction algorithms are actively being studied at low doses to solve dose problems and reduce noise at the same time. Wiener filter and median filter have been widely used, with the disadvantages of poor noise reduction efficiency and loss of much information about imaging boundary. The purpose of this study is to apply total variation (TV) algorithm to skull X-ray imaging that can compensate for the problems of previous noise reduction efficiency to assess quantitatively and compare them. For this study, skull X-ray imaging is obtained using various kVp and mAs using the skull phantom using the X-ray device of Siemens. In addition, contrast to noise ratio (CNR) and coefficient of variation (COV) are compared and measured when noisy image, median filter, Wiener filter and TV algorithm were applied to each phantom imaging. Experiments showed that when TV algorithms were applied, CNR and COV characteristics were excellent under all conditions. In conclusion, we've been able to see if we can use TV algorithm to improve image quality and CNR could be seen to increase due to the decrease in noise as the amount of increased mAs. On the other hand, COV decreased as the amount of increased mAs, and when kVp increased, noise was reduced and the transmittance was increased, so COV was reduced.

• Journal of Convergence for Information Technology
• /
• 제11권7호
• /
• pp.189-194
• /
• 2021

The purpose of this study was to investigate the effect of reducing noise and streaking artefacts by applying the Boost3D algorithm in the case of noise and streaking artifacts generated during computed tomography of the shoulder joint. A phantom study using a thoracic phantom including shoulder joint and clinical evaluation were conducted through shoulder joint images of 35 patients who underwent computed tomography of the shoulder joint from September 2020 to October 2020. The evaluation was divided into groups before and after the application of the Boost3D algorithm, and the noise values, signal to noise ratio, and mean to standard deviation ratio values were analyzed. Both noise values and mean to standard deviation ratio values analyzed in phantom image evaluation and clinical image evaluation were statistically significantly lower in the group after Boost3D was applied (p<0.05). Through this study, it was found that noise and streak artifacts were reduced through the application of Boost3D, and the mean to standard deviation ratio was high, which can be judged as an excellent image. If the Boost3D algorithm is used for computed tomography of the shoulder joint, it is thought that excellent images can be obtained with reduced noise and streaking artifacts that may occur in the shoulder joint area.

• The Journal of the Korea Contents Association
• /
• 제17권5호
• /
• pp.491-498
• /
• 2017

This study is to determine the significant effect of white and pink noise on brain. The brain synchronization has been analyzed under the condition of non-noise, white nose and pink noise(male 10, female 10, mean age $23.3{\pm}2.14$). As a result of analysis, pink noise stimulus, alpha, low beta band, and high beta band were significantly decreased than non-noise and white noise. In addition, these brain response pattern significantly increased at frontal lobe and temporal lobe, and dominated on the right hemisphere. This result is considered to be useful of sound design in driving quality of human life on the basis of neuroscience.

• Journal of the Korean Society of Radiology
• /
• 제11권1호
• /
• pp.63-68
• /
• 2017

To reduce the exposure dose in head CT, the use of low tube voltage is required. However, increasing noise may cause errors in the second data processing. In this study, we proposed a method to reduce noise by using low tube voltage. Experimental results show that the noise level is high at 100kVp and lowest at 140 kVp. The dose was lower at 100 kVp and higher at 140 kVp. As a result of applying the wavelet according to the threshold value, the noise value in the wavelet Th30 decreased to 4.51. Using the parameter condition(100 kVp, rotation time 0.5 sec, dose: 40.64 mGy) and the wavelet Th 30, the dose reduction of 65.3% was possible. We believe that applying the proposed method to head CT images will help to patient safety and interpret accurate information.