• Title/Summary/Keyword: Network reduction

Search Result 1,414, Processing Time 0.028 seconds

Camera noise reduction in the low illumination conditions using convolutional network (컨벌루션 네트워크를 이용한 저조도 환경 카메라 잡음 제거)

  • Park, Gu-Yong;Ahn, Byeong-Yong;Cho, Nam-ik
    • Proceedings of the Korean Society of Broadcast Engineers Conference
    • /
    • 2017.06a
    • /
    • pp.163-165
    • /
    • 2017
  • 본 논문에서는 카메라 잡음 제거에 딥 러닝 알고리즘을 적용하는 연구를 진행하였다. 합성된 가우시언 잡음에 대하여 좋은 잡음 제거 성능을 보이는 DnCNN(Denoising Convolutional Network)를 이용하여 카메라 잡음을 제거하는 학습과 실험을 진행하였으며, 기준 실험으로는 RGB 색공간의 3채널 모두에 대하여 학습한 신경망(Neural Network)을 사용하였고, 본 논문의 실험에서는 그레이 이미지에 대하여 학습한 신경망을 사용하였다. 신경망의 평가를 위하여 딥 러닝 알고리즘 입력 이미지를 RGB 색공간(RGB Color Space)과 YCbCr 색공간(YCbCr Color Space) 2가지 색공간으로 표현하여 사용하였고, 입력 이미지에 노이즈를 첨가하기 위해 가우시안 노이즈(Gaussian Noise)를 이용하였다. 또한 가우시안 잡음과 다른 성질을 갖는 실제 카메라 잡음에 대해서도 학습과 테스트를 진행하였다.

  • PDF

Generative Adversarial Network based CNN model for artifact reduction on HEVC-encoded video (HEVC 비디오 영상 압축 왜곡 제거를 위한 Generative Adversarial Network 적용 기법)

  • Jeon, Jin;Kim, Munchurl
    • Proceedings of the Korean Society of Broadcast Engineers Conference
    • /
    • 2017.06a
    • /
    • pp.192-193
    • /
    • 2017
  • 본 논문에서는 비디오 영상 압축 왜곡 제거를 위해 Generative Adversarial Network (GAN)을 적용한 컨벌루션 뉴럴 네트워크 (CNN) 모델을 제안한다. GAN 모델의 생성 모델 (Generator)은 노이즈가 아닌 High Efficiency Video Coding (HEVC)로 압축된 영상을 입력 받은 뒤, 압축 왜곡이 제거된 영상을 출력하며, 분류 모델 (Discriminator)은 원본 영상과 압축된 영상을 입력 받은 뒤, 원본 영상과 압축 왜곡이 포함된 압축된 영상을 분류한다. 분류 모델은 5 개 층을 쌓은 컨벌루션 뉴럴 네트워크 구조를 사용하였고, 생성 모델은 5 개 층을 쌓은 SRCNN 구조와 VDSR 구조를 기반으로 한 두 개의 모델을 이용한 실험을 통해 얻은 결과를 비교하였다. 비디오 영상 압축 왜곡 제거 실험을 위해 원본 비디오 영상을 HEVC 을 이용하여 2Mbps, 4Mbps 로 압축된 영상을 사용하였으며, 압축된 영상 대비 왜곡이 제거된 영상을 얻을 수 있었다.

  • PDF

Realization of a Parallel Network System for Image Processing Techniques (영상 처리 기법을 위한 병렬화 네트워크 시스템의 구성)

  • 서원찬;조강현;김우열
    • Journal of Institute of Control, Robotics and Systems
    • /
    • v.6 no.6
    • /
    • pp.492-499
    • /
    • 2000
  • In this paper, realization techniques of the parallel processing and the parallel network system for image processing are described. The parallel image processing system is constructed by the characterization of image processing and processor. Several problems are solved to achieve effective parallel processing and processor networking with the particular properties of image processing, which are reduction of communication quantity, equalization of load and delay depreciation on communication. A parallel image input device is developed for the flexible networking of parallel image processing. An abnormal region detection algorithm which is the basic function in machine vision is applied to evaluate the constructed parallel image processing system. The performance and effectiveness of the system are confirmed by experiments.

  • PDF

Network Reduction Method for Power System Transient Analysis ; Time-Domain Formulation Based On The Network Function (과도 상태 해석을 위한 계통 축약법 ; 계통 함수를 이용한 시간 영역 해석법)

  • Hong, J.H.;Kang, Y.C.;Cho, K.R.;Park, J.K.
    • Proceedings of the KIEE Conference
    • /
    • 1991.07a
    • /
    • pp.417-421
    • /
    • 1991
  • In electromagnetic transient analysis, complex transmission system should be modelled in detail. But in large system, this full representation of power transmission system has a big burden in many sides such as computation time, modelling efforts, etc. It is very required, therefore, in electromagnetic transients studies to represent parts of a complete system in a reduced or an equivalent form. This paper develops the method from which system equivalents may be derived. The suggested method is of an essentially transient form, and allows travelling wave interaction between the equivalent and explicit transmission network to be modelled.

  • PDF

Effect of Hot Forging on the Hardness and Toughness of Ultra High Carbon Low Alloy Steel (초 고 탄소 저합금강의 경도와 인성에 미치는 열간단조의 영향)

  • Kim, Jong-Beak;Kang, Chang-Yong
    • Journal of Power System Engineering
    • /
    • v.17 no.6
    • /
    • pp.115-121
    • /
    • 2013
  • This study was carried out to investigate the effect of hot forging on the hardness and impact value of ultra high carbon low alloy steel. With increasing hot forging ratio, thickness of the network and acicular proeutectoid cementite decreased, and than were broken up into particle shapes, when the forging ratio was 80%, the network and acicular shape of the as-cast state disappeared. Interlamellar spacing and the thickness of eutectoid cementite decreased with increasing forging ratio, and were broken up into particle shapes, which then became spheroidized. With increasing hot forging ratio, hardness, tensile strength, elongation and impact value were not changed up 50%, and then hardness rapidly decreased, while impact value rapidly increased. Hardness and impact value was greatly affected by the disappeared of network and acicular shape of proeutectoid cementite, and became particle shape than thickness reduction of proeutectoid and eutectoid cementite.

A Study on the Fluid Network Analysis for the LPG Supply System of the Gaseous Fuel Injection Type (LPG 가스분사 방식 연료공급시스템의 관로 유동해석에 관한 연구)

  • Yun, Jeong-Eui;Kim, Myung-Hwan;Nam, Hyeon-Sik;Jeong, Tae-Hyuung
    • Transactions of the Korean Society of Automotive Engineers
    • /
    • v.15 no.2
    • /
    • pp.35-40
    • /
    • 2007
  • The gaseous fuel injection (GFI) type in LPG fuel supply system has more advantage than the liquified fuel injection type from the viewpoint of durability and cost reduction. But in GFI system, to control pressure and temperature of gaseous fuel is needed to get precision fuel metering for the compressible characteristic of gaseous fuel. In this study, the effects of pressure and temperature on the fuel metering was simulated by commercial flow network analysis package, Flowmaster. And the fuel composition effects on the fuel metering were also studied to figure out the fuel metering characteristics.

Probabilistic Broadcasting Based on Selfishness and Additional Coverage in MANETs

  • Kim, Jae-Soo;Kim, Jeong-Hong
    • Journal of information and communication convergence engineering
    • /
    • v.10 no.4
    • /
    • pp.329-336
    • /
    • 2012
  • For designing broadcast protocols in mobile ad hoc networks (MANETs), one of the important goals is to reduce the rebroadcast packets redundancy while reaching all the nodes in network. In this paper, we propose a probabilistic broadcasting mechanism based on selfishness and additional coverage in MANETs. Our approach dynamically adjusts the rebroadcast probability according to the extra covered area and number of neighbor nodes. By these two factors, mobile hosts can be classified into three groups: normal, low selfishness, and high selfishness groups. The nodes in the normal group forward packets for other nodes with high probability, whereas the nodes in the low selfishness group rebroadcast packets with low probability and the nodes in the high selfishness group do not rebroadcast packets. We compared our approach with simple flooding and the fixed probabilistic approach. The simulation results show that the proposed schemes can significantly reduce the number of retransmissions by up to 40% compared simple flooding and fixed probabilistic scheme without significant reduction in the network reachability and end-to-end packet delay.

Smart Control System Using Fuzzy and Neural Network Prediction System

  • Kim, Tae Yeun;Bae, Sang Hyun
    • Journal of Integrative Natural Science
    • /
    • v.12 no.4
    • /
    • pp.105-115
    • /
    • 2019
  • In this paper, a prediction system is proposed to control the brightness of smart street lamps by predicting the moving path through the reduction of consumption power and information of pedestrian's past moving direction while meeting the function of existing smart street lamps. The brightness of smart street lamps is adjusted by utilizing the walk tracking vector and soft hand-off characteristics obtained through the motion sensing sensor of smart street lamps. In addition, the motion vector is used to analyze and predict the pedestrian path, and the GPU is used for high-speed computation. Pedestrians were detected using adaptive Gaussian mixing, weighted difference imaging, and motion vectors, and motions of pedestrians were analyzed using the extracted motion vectors. The preprocessing process using linear interpolation is performed to improve the performance of the proposed prediction system. Fuzzy prediction system and neural network prediction system are designed in parallel to improve efficiency and rough set is used for error correction.

An Operational Strategy for Inventory Control of Networked Regional Distribution Centers (지역통합 네트워크관리하의 재고통제 운용전략에 관한 연구)

  • Kim, Byeong-Chan;Choi, Jin-Yeong
    • Journal of Korean Society of Industrial and Systems Engineering
    • /
    • v.31 no.3
    • /
    • pp.110-116
    • /
    • 2008
  • Operational strategy for inventory control in the distribution system has been given attention. If an individual enterprise implements the strategy, it is not easy to gain scale merits because of limited quantity or burden of inventory. In this study, we propose an operational strategy for inventory control that considers managerial integration of regional distribution centers (RDCs) and present a model of it. In a network of several RDCs, they could share inventory information and supply parts for others in case of an inventory shortage. And a numerical example of the network is illustrated, which compares two operational strategies, integration management of RDCs and individual management of them. The result shows total cost reduction in the strategy of integration management through the efficient inventory control of multi-echelon distribution.

The Competitive Time Guarantee Decisions Via Continuous Approximation of Logistics Systems (연속적 근사법에 의한 물류시스템의 경쟁적 시간보장 의사결정 최적화에 관한 연구)

  • Kim, Hyoungtae
    • Journal of Korean Society of Industrial and Systems Engineering
    • /
    • v.37 no.3
    • /
    • pp.64-74
    • /
    • 2014
  • We show how a supplier can peg cost measures to the reliability of his time guarantees via the penalty costs considered in the framework. The framework also enables us to study the connections between the logistics network and the market. In this context, we show that even when the market base increases significantly, the supplier can still use the logistics network designed to satisfy lower demand density, with only a marginal reduction in profit. Finally we show how the framework is useful to evaluate and compare various logistics system improvement strategies. The supplier can then easily choose the improvement strategy that increases his profit with the minimal increase in his logistics costs.