• Proceedings of the KIEE Conference
• /
• 1995.07a
• /
• pp.15-17
• /
• 1995

This paper presents compression algorithm of a system matrix for electromagnetic analysis by the finite element method. Generally the solution of the finite element analysis is the more accurate the more number of nodes. The memory of a computer limit to number of nodes. Therefore it is needed the technique of compress the system matrix. This algorithm is useful to handle non-zero-terms that can be generated during the application of boundary condition.

• Proceedings of the KIEE Conference
• /
• 2003.11c
• /
• pp.1010-1013
• /
• 2003

This paper describes an algorithm for determining robot location using appearance-based paradigm. This algorithm compress the image set using PCA(principal component analysis) to obtain a low-dimensional subspace, called the eigenspace, and it makes a manifold that represent a continuous-appearance function. To determine robot location, given an unknown input image, the recognition system first projects the image to eigenspace. Neural network use coefficients of the eigenspace to estimate the location of the mobile robot. The algorithm has been implemented and tested on a mobile robot system. In several trials it computes location accurately.

• Proceedings of the Korea Institute of Convergence Signal Processing
• /
• 2000.08a
• /
• pp.253-256
• /
• 2000

In this paper, we propose the lossless image compression algorithm using the integer wavelet transform. Recently, the S＋P transform is widely used and computed with only integer addition and bit-shift operations, but not proper to remove the correlation of smooth images. then we compare the Harr wavelet of the S＋P transform with various integer coefficient filter banks and apply 9/7 ICFB to the wavelet transform. In addition, we propose a entropy-coding method that exploits the multiresolution structure and the feedback of the prediction error, and can efficiently compress the transformed image for progressive transmission. Simulation results are compared to the compression ratio using the S+P transform with different types of images.

• Aerospace Engineering and Technology
• /
• v.7 no.1
• /
• pp.216-222
• /
• 2008

Purpose of this research is estimation of water depth by hyperspectral remote sensing in area that access of ship is difficult. This research used EO-l Hyperion satellite imagery. Atmospheric and geometric correction is executed. Compress of band used MNF transforms. Diffuse Attenuation Coefficient of target area is decided in imagery for water depth estimation. Determination of Emdmember in pixel is using Linear Spectral Unmixing techniques. Water depth estimated using this result.

• Journal of Korea Multimedia Society
• /
• v.22 no.1
• /
• pp.10-17
• /
• 2019

In High Dynamic Range (HDR) image processing, tone mapping is the process to compress an input image into a Low Dynamic Range (LDR) image. In most cases, the reason that detail preservation is prior to take over tone mapping is that the dynamic range is significantly different between input and output images. In the case of iCAM06, details are separated by using a bilateral filter, however, it causes noise amplification at the dim surround region. Thus, we suggest that the detail signal, which is separated from the bilateral filter, is combined with the base signal after an adaptive Kalman filter is applied according to the local standard deviation. We confirmed that the proposed method enhances the HDR images quality by checking the noise reduction in a dim surround region.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.67 no.12
• /
• pp.1665-1670
• /
• 2018

A convolutional neural network (CNN), which is one of the deep learning models, has been very successful in a variety of computer vision tasks. Filters of a CNN are automatically generated, however, they can be further optimized since there exist the possibility of existing redundant and less important features. Therefore, the aim of this paper is a filter reduction to accelerate and compress CNN models. Evolutionary algorithms is adopted to remove the unnecessary filters in order to minimize the parameters of CNN networks while maintaining a good performance of classification. We demonstrate the proposed filter reduction methods performing experiments on CIFAR10 data based on the classification performance. The comparison for three approaches is analysed and the outlook for the potential next steps is suggested.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2019.06a
• /
• pp.84-85
• /
• 2019

MPEG-NNR (Compression of Neural Network for Multimedia Content Description and Analysis) aims to define a compressed and interoperable representation of trained neural networks. In this paper, we present a low-rank approximation to compress a CNN used for image compression, which is one of MPEG-NNR use cases. In the presented method, the low-rank approximation decomposes one 2D kernel matrix of weights into two 1D kernel matrix values in each convolution layer to reduce the data amount of weights. The evaluation results show that the model size of the original CNN is reduced to half as well as the inference runtime is reduced up to about 30% with negligible loss in PSNR.

• The Bulletin of The Korean Astronomical Society
• /
• v.43 no.2
• /
• pp.54.1-54.1
• /
• 2018

Recent observational studies suggest that the environmental effects can shape the evolution of galaxies in clusters. In an attempt to better understand this process, we perform idealized radiation-hydrodynamic simulations of RAM pressure stripping on star-forming galaxies using RAMSES-RT. We find that extended HI disks are easily stripped by moderate ICM winds, while there is no significant decrease in the total mass of molecular gas. RAM pressure tends to compress the molecular gas, leading to enhanced star formation especially when the gaseous disk is hit by edge-on winds. On the other hand, strong ICM winds that are expected to operate at the centre of clusters strip both HI and molecular gas from the galaxy. Interestingly, we find that the strong ICM winds can induce the formation of relatively dense (~1H/cc) HI gas clouds at a distance from the disk.

• Journal of Hydrogen and New Energy
• /
• v.31 no.6
• /
• pp.578-586
• /
• 2020

There is growing interest worldwide in a hydrogen economy that uses hydrogen as an energy medium instead of hydrocarbon-based fossil fuels as a way to combat climate change. Since hydrogen has a very low energy density per unit volume at room temperature, hydrogen must be compressed and stored in order to use as an energy carrier. There are mechanical and non-mechanical methods for compressing hydrogen. The mechanical method has disadvantages such as high energy consumption, durability problems of moving parts, hydrogen contamination by lubricants, and noise. Among the non-mechanical compression methods, electrochemical compression consumes less energy and can compress hydrogen with high purity. In this paper, research trends are reviewed, focusing on research papers on electrochemical hydrogen compression technology, and future research directions are suggested.

• Journal of the Semiconductor & Display Technology
• /
• v.21 no.4
• /
• pp.156-162
• /
• 2022

Image compression plays an important role in encoding and improving various forms of images in the digital era. Recent researches have focused on the principle of deep learning as one of the most exciting machine learning methods to show that it is good scheme to analyze, classify and compress images. Various neural networks are able to adapt for image compressions, such as deep neural networks, artificial neural networks, recurrent neural networks and convolution neural networks. In this review paper, we discussed how to apply the rule of deep learning to obtain better image compression with high accuracy, low loss-ness and high visibility of the image. For those results in performance, deep learning methods are required on justified manner with distinct analysis.