• The Journal of the Acoustical Society of Korea
• /
• v.14 no.1
• /
• pp.13-21
• /
• 1995

In this paper, we improve an image classifier algorithm based on neural network learning. It consists of two steps. The first is input pattern generation and the second, the global neural network implementation using an improved back-propagation algorithm. The feature vector for pattern recognition consists of the codebook data obtained from self-organization feature map learning. It decreases the input neuron number as well as the computational cost. The global neural network algorithm which is used in classifier inserts a control part and an address memory part to the back-propagation algorithm to control weights and unit-offsets. The simulation results show that it does not fall into the local minima and can implement easily the large-scale neural network. And it decreases largely the learning time.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.528-529
• /
• 2023

Due to vast points and irregular structure, labeling full points in large-scale point clouds is highly tedious and time-consuming. To resolve this issue, we propose a novel point-based transformer network in weakly-supervised semantic segmentation, which only needs 0.1% point annotations. Our network introduces general local features, representing global factors from different neighborhoods based on their order positions. Then, we share query point weights to local features through point attention to reinforce impacts, which are essential in determining sparse point labels. Geometric encoding is introduced to balance query point impact and remind point position during training. As a result, one point in specific local areas can obtain global features from corresponding ones in other neighborhoods and reinforce from its query points. Experimental results on benchmark large-scale point clouds demonstrate our proposed network's state-of-the-art performance.

• Journal of Internet Computing and Services
• /
• v.17 no.3
• /
• pp.25-32
• /
• 2016

In this paper, we propose a method that is tracking an object in real time using particle filter and the observation model with edge. First of all, the proposed method defines the object to be tracked in the initial frame. Then, it generates the edge observation model for the object to be tracked and a set of particles. It calculates the weight by comparing the average of the middle distance in eight-way of particle filter edge model with that in edge observation model, and then updates the weight with the calculated value. After resampling particles using the updated weights, it estimates the current location of the tracked object. Finally, this paper demonstrates the performance of the stable tracking through comparison with the existing method by using a number of experimental data.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.7
• /
• pp.2769-2784
• /
• 2020

This paper aims to compare the scenic images of four ancient Chinese cities including Lijiang, Pingyao, Huizhou and Langzhong, so as to provide specific development strategies for the ancient cities. In this paper, the ancient cities' scenic images are divided into three sub-indexes and eight evaluation dimensions. Based on this, the study first uses Python software to collect tourists' online comments on the four ancient cities. Then, the social network analysis method is used to build a high-frequency keywords matrix of tourist comments and the R language is used to generate a visual network graph. After this, the entropy weight method is used to determine the weights and values of eight evaluation dimensions. Finally, the tourists' overall satisfaction indexes of the four ancient cities are calculated accordingly. The results show that (1) the overall satisfaction of Lijiang is the highest, while that of Huizhou is the lowest; (2) from the weight of each evaluation dimension, it can be seen that tourists care more about the national culture and historical culture; (3) from tourists' satisfaction index on each evaluation dimension of the four ancient cities, we can find that the four ancient cities has their own advantages and disadvantages in tourism development. (4) local tourism-related institutions should strengthen their advantages and improve their deficiencies so as to enhance tourists' overall image of the ancient city.

• Journal of the Institute of Convergence Signal Processing
• /
• v.12 no.2
• /
• pp.113-118
• /
• 2011

In this paper, we develop a speaker gender classification technique using collaborative sensor fusion for use in a wireless sensor network. The distributed sensor nodes remove the unwanted input data using the BER(Band Energy Ration) based voice activity detection, process only the relevant data, and transmit the hard labeled decisions to the fusion center where a global decision fusion is carried out. This takes advantages of power consumption and network resource management. The Bayesian sensor fusion and the global weighting decision fusion methods are proposed to achieve the gender classification. As the number of the sensor nodes varies, the Bayesian sensor fusion yields the best classification accuracy using the optimal operating points of the ROC(Receiver Operating Characteristic) curves_ For the weights used in the global decision fusion, the BER and MCL(Mutual Confidence Level) are employed to effectively combined at the fusion center. The simulation results show that as the number of the sensor nodes increases, the classification accuracy was even more improved in the low SNR(Signal to Noise Ration) condition.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.30 no.6
• /
• pp.1271-1278
• /
• 2020

Deep learning-based profiling side-channel analysis is a powerful analysis method that utilizes the neural network to profile the relationship between the side-channel information and the intermediate value. Since the neural network interprets each point of the signal in a different dimension, jitter makes it much hard that the neural network with dimension-wise weights learns the relationship. This paper shows that replacing the fully-connected layer of the traditional CNN (Convolutional Neural Network) with global average pooling (GAP) allows us to design the inherently robust neural network inherently for jitter. We experimented with the ChipWhisperer-Lite board to demonstrate the proposed method: as a result, the validation accuracy of the CNN with a fully-connected layer was only up to 1.4%; contrastively, the validation accuracy of the CNN with GAP was very high at up to 41.7%.

• Transactions of Materials Processing
• /
• v.22 no.1
• /
• pp.42-47
• /
• 2013

The bicycle is one of the most important eco-friendly transport methods which can mitigate global warming. The portability of a bike on public transportation systems is essential for the wide spread use of bicycles by people in urban environments. In this study, a lightweight urban folding bike was developed with careful consideration of the association with public transport. A folding frame using a moving slide link mechanism made from AL6061 is proposed. Numerical analysis was conducted to evaluate structural safety of the bike in both vertical and pedal loading tests. The proposed urban folding bicycle weights only 10kg and summation of its width, length, and height in the folded configuration is under 158cm.

• International Journal of Aeronautical and Space Sciences
• /
• v.11 no.3
• /
• pp.155-166
• /
• 2010

I review research on self-sensing and structural optimizations of laminated carbon/epoxy composites in Japan. Self-sensing is one of the multiple functions of composites; i.e., carbon fiber is used as a sensor as well as reinforcement. I present a controversial issue in self-sensing and detail research results. Structural optimization of laminated CFRP composites is indispensable in reducing the weights of modern aerospace structural components. I present a modified efficient global search method using the multi-objective genetic algorithm and fractal branch and bound method. My group has focused its research on these subjects and our research results are presented here.

• Journal of Korean Society for Quality Management
• /
• v.31 no.2
• /
• pp.207-219
• /
• 2003

The future will see all industries become technology-driven in the competitive global marketplace. Venture firms with deep technological roots and innovation strategies have some advantages. In this situation, existing methods for technology assessment are not enough to evaluate their relative technology competitiveness. Therefore, a more useful and comprehensive approach is needed to obtain the desired outcomes for measuring the relative competitiveness of technology effectively. In this research, I applied factor analysis, which is methodology to be capable of determining the common factors and associated weights reasonably, to the development of the technology competitiveness indexes.

• Journal of Energy Engineering
• /
• v.27 no.2
• /
• pp.70-80
• /
• 2018

Natural organic matter (NOM) is among the most common pollutant in underground and surface waters. It comprises of humic substances which contains anionic macromolecules such as aliphatic and aromatic compounds of a wide range of molecular weights along with carboxylic, phenolic functional groups. Although the concentration of NOM in potable water usually lies in the range of 1-10 ppm. Conventional treatment technologies are facing challenge in removing NOM effectively. The main issues are concentrated to low efficiency, membrane fouling, and harmful by-product formation. Ion-exchangers can be considered as an efficient and economic pretreatment technology for the removal of NOM. It not only consumes less time for pretreatment but also resist formation of trihalomethanes (THMs), an unwanted harmful by-product. This article provides a comprehensive review of ion exchange processes for the removal of NOM.