• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2022.11a
• /
• pp.4-7
• /
• 2022

기존 합성곱 신경망 기반의 잡음 제거 네트워크들은 학습을 위한 noisy-clean 데이터 쌍을 필요로 한다. 하지만 실제 카메라 잡음의 경우, 잡음에 대한 깨끗한 원본 영상을 얻는 것은 불가능하거나 많은 비용이 소모된다. 따라서 이러한 방법을 해결하기 위하여 원본 영상 없이 잡음 영상만으로만 잡음 제거 네트워크를 학습하는 방법들이 제안되어왔다. 그 중 카메라 잡음 영상을 처리하기 위한 대표적인 방법으로 학습과 추론에서 비대칭적인 downsampling을 사용하는 AP-BSN이 제안되었다. 본 논문에서는 Functional neural network를 AP-BSN 알고리즘에 적용하여 다양한 downsampling ratio에 대응되는 하나의 네트워크를 학습하였다. 이를 통해 기존 hyperparameter로 사용되던 downsampling ratio에 대한 결과를 하나의 네트워크에서 분석 및 확인하였다. 또한 해당 파라미터를 조절함으로써 다양한 잡음 제거 후보들을 추출하고 사용자가 원하는 잡음 제거 정도를 조정할 수 있도록 하였다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2021.06a
• /
• pp.176-179
• /
• 2021

최근 딥 러닝 기법의 하나인 합성곱 신경망(Convolutional Neural Network, CNN)은 영상 잡음(Noise) 제거 분야에서 전통적인 기법보다 좋은 성능을 나타내고 있지만 학습하는 과정에서 영상 내 디테일한 부분이 손실될 수 있다. 본 논문에서는 웨이블릿 변환(Wavelet Transform)을 기반으로 영상 내 디테일 정보도 같이 학습하여 영상 디테일을 향상하는 잡음 제거 합성곱 신경망 네트워크를 제안한다. 제안하는 네트워크는 디테일 향상 서브 네트워크(Detail Enhancement Subnetwork)와 영상 잡음 추출 서브 네트워크(Noise Extraction Subnetwork)를 이용하게 된다. 실험을 통해 제안하는 방법은 기존 알고리듬보다 디테일 손실 문제를 효과적으로 해결할 수 있었고 객관적 품질 평가인 PSNR(Peak Signal-to-Noise Ratio)와 주관적 품질 비교에서 모두 우수한 결과가 나온 것을 확인하였다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.11a
• /
• pp.960-962
• /
• 2021

Monte Carlo 렌더링은 모든 빛을 광원에서부터 추적하는 것 대신, 몇 개의 빛의 경로만을 추적해서 이들의 평균으로 화소값을 정해 이미지를 만드는 방법이다. 여기서 추적하는 빛이 많다면 이미지가 사실적으로 만들어질 수 있지만 연산량이 증가한다. 따라서 적은 빛의 경로를 추적하여 렌더링을 수행하여 이미지를 만들고, 노이즈를 제거해서 많은 양의 빛을 추적하여 렌더링을 한 이미지와 유사하게 만들려는 연구가 많이 진행되고 있다. 그러나 이러한 연구들은 많은 연산량을 요구하기 때문에 고성능의 기기 사양을 요구한다. 따라서 본 연구에서는 저사양의 기기에서 활용할 수 있도록 Harris corner 검출법과 median filtering을 활용한 렌더링 이미지 노이즈 제거 연구를 수행했다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2022.06a
• /
• pp.217-220
• /
• 2022

본 논문에서는 CCTV 영상 화질을 향상하고 해상도를 높이기 위해 딥 러닝(Deep Learning)을 이용하여 잡음 제거(Denoising) 와 초해상도(Super-resolution) 작업을 수행한다. 데이터 증강(Data Augmentation)을 통한 초해상도 성능 향상을 위해서 잡음 제거 네트워크의 출력 영상을 초해상도 네트워크의 입력으로 사용하는 순차적 작업을 사용한다. 또한 딥 러닝을 이용한 영상처리에서 발생하는 평균 밝기 오차 문제를 해결하기 위한 손실함수(Loss Function)와 두 가지 이상의 순차적인 딥 러닝 작업에서 발생하는 문제점을 극복하기 위한 손실함수를 제안한다. 제안하는 손실함수는 네트워크의 출력 영상과 타겟 영상의 밝기 오차를 줄이는 것이 가능하고, 순차적 작업에서 보다 정확한 모델 성능 판단이 가능하다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.18 no.2
• /
• pp.475-481
• /
• 2014

As the society develops to advanced digital information times, many studies are underway about digital video processing technology areas such as image restoration. There are typical methods to restore the image which have been damaged by the impulse noise like SM(standard median) filter and CWM(center weighted median) filter. These filters show excellent noise reduction capabilities in low noise density areas, but in high noise density areas, noise reduction capabilities are not sufficient. In this paper, in order to restore the degraded images in impulse(Salt & Pepper) noise environment, the image restoration filter algorithm was suggested which expands and subdivide the mask focusing on damaged pixels. And to demonstrate the superiority of the proposed algorithm used PSNR (peak signal to noise ratio) as the standard of judgement.

• Journal of the Korean Society of Radiology
• /
• v.12 no.6
• /
• pp.707-712
• /
• 2018

The number of CT examinations is increasing, and radiation exposure is also increasing. repeated tests can affect the lens and thyroid. In hospitals, there is a tendency to lack interest in major long-term radiation exposure compared to the interest in increasing image information and image quality with head CT. In this study, we analyzed the improvement of image quality by proposed method to the noisy CT images. The proposed denoising method total variance optimization only for the impulsive noise candidate pixels. Experimental results show that edge information is well preserved and impulse noise can be effectively removed. and worked very well for the images according to tube voltage and rotation time. applied to the clinical setting, it can be used as the lowest exposure condition without worrying about the image quality and it will be helpful for the CT application.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.2
• /
• pp.311-326
• /
• 2024

The rapid development of neural network technology promotes the neural network model driven by big data to overcome the texture effect of complex objects. Due to the limitations in complex scenes, it is necessary to establish custom template matching and apply it to the research of many fields of computational vision technology. The dependence on high-quality small label sample database data is not very strong, and the machine learning system of deep feature connection to complete the task of texture effect inference and speculation is relatively poor. The style transfer algorithm based on neural network collects and preserves the data of patterns, extracts and modernizes their features. Through the algorithm model, it is easier to present the texture color of patterns and display them digitally. In this paper, according to the texture effect reasoning of custom template matching, the 3D visualization of the target is transformed into a 3D model. The high similarity between the scene to be inferred and the user-defined template is calculated by the user-defined template of the multi-dimensional external feature label. The convolutional neural network is adopted to optimize the external area of the object to improve the sampling quality and computational performance of the sample pyramid structure. The results indicate that the proposed algorithm can accurately capture the significant target, achieve more ablation noise, and improve the visualization results. The proposed deep convolutional neural network optimization algorithm has good rapidity, data accuracy and robustness. The proposed algorithm can adapt to the calculation of more task scenes, display the redundant vision-related information of image conversion, enhance the powerful computing power, and further improve the computational efficiency and accuracy of convolutional networks, which has a high research significance for the study of image information conversion.

• Journal of the Korean Society of Radiology
• /
• v.17 no.2
• /
• pp.207-213
• /
• 2023

Computed tomography (CT) images are used as the basis for proton Bragg peak position estimation and treatment plan simulation. During the Hounsfield Unit (HU) based proton stopping power ratio (SPR) estimation, small differences in the patient's density and elemental composition lead to uncertainty in the Bragg peak positions along the path of the proton beam. In this study, we investigated the potential of dual-energy computed tomography image-based proton SPRs prediction accuracy to reduce the uncertainty of Bragg peak position prediction. Single- and dual-energy images of an electron density phantom (CIRS Model 062M electron density phantom, CIRS Inc., Norfolk, VA, USA) were acquired using a computed tomography system (Somatom Definition AS, Siemens Health Care, Forchheim, Germany) to estimate the SPRs of the proton beam. To validate the method, it was compared to the SPRs estimated from standard data provided by the National Institute of Standards and Technology (NIST). The results show that the dual-energy image-based method has the potential to improve accuracy in predicting the SPRs of proton beams, and it is expected that further improvements in predicting the position of the proton's Bragg peak will be possible if a wider variety of substitutes with different densities and elemental compositions of the human body are used to predict the SPRs.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.19 no.4
• /
• pp.965-972
• /
• 2015

Along with the rapid development in digital times, image media are being used in internet, computer and digital camera. But image deterioration occurs due to various exterior reasons in the procedures of acquisition, processing, transmission and recording of digital image and major reason is noise. Therefore in order to remove salt & pepper noise, this study suggested the algorithm which replaces the noise to original pixel in case of non-noise, and processes the noise with Lagrange interpolation method in case of noise. In case high density noise was added and the noise could not be removed, noise characteristics were improved by processing the noises repeatedly. And for objective judgment, this method was compared with existing methods and PSNR(peak signal to noise ratio) was used as judgment standard.

• Journal of IKEEE
• /
• v.23 no.4
• /
• pp.1366-1372
• /
• 2019

In this paper, we designed a revised neural network to remove the Monte Carlo Rendering noise contained in the ray tracing graphics. The Monte Carlo Rendering is the best way to enhance the graphic's realism, but because of the need to calculate more than thousands of light effects per pixel, rendering processing time has increased rapidly, causing a major problem with real-time processing. To improve this problem, the number of light used in pixels is reduced, where rendering noise occurs and various studies have been conducted to eliminate this noise. In this paper, a deep learning is used to remove rendering noise, especially by separating the rendering image into diffuse and specular light, so that the structure of the dual neural network is designed. As a result, the dual neural network improved by an average of 0.58 db for 64 test images based on PSNR, and 99.22% less light compared to reference image, enabling real-time race-tracing rendering.