• Journal of Internet of Things and Convergence
• /
• v.9 no.1
• /
• pp.9-17
• /
• 2023

Deep learning with large amount of computations is difficult to implement on micro-sized IoT devices or moblie devices. Recently, lightweight deep learning technologies have been introduced to make sure that deep learning can be implemented even on small devices by reducing the amount of computation of the model. Quantization is one of lightweight techniques that can be efficiently used to reduce the memory and size of the model by expressing parameter values with continuous distribution as discrete values of fixed bits. However, the accuracy of the model is reduced due to discrete value representation in quantization. In this paper, we introduce various quantization techniques to correct the accuracy. We selected APoT and EWGS from existing quantization techniques, and comparatively analyzed the results through experimentations The selected techniques were trained and tested with CIFAR-10 or CIFAR-100 datasets in the ResNet model. We found out problems with them through experimental results analysis and presented directions for future research.

• Earthquakes and Structures
• /
• v.18 no.3
• /
• pp.337-348
• /
• 2020

Experimental testing has been considered as one of the most straightforward approaches to realize the structural behavior for earthquake engineering studies. Recently, novel and advanced experimental techniques, which combine numerical simulation with experimental testing, have been developed and applied to structural testing practically. However, researchers have to take the risk of damaging specimens or facilities during the process of developing and validating new experimental methods. In view of this, a small-scale structural laboratory has been designed and constructed in order to verify the effectiveness of newly developed experimental technique before it is applied to large-scale testing for safety concerns in this paper. Two orthogonal steel reaction walls and one steel T-slotted reaction floor are designed and analyzed. Accordingly, a large variety of experimental setups can be completed by installing servo-hydraulic actuators and fixtures depending on different research purposes. Meanwhile, a state-of-the-art digital controller and multiple real-time computation machines are allocated. The integration of hardware and software interfaces provides the feasibility and flexibility of developing novel experimental methods that used to be difficult to complete in conventional structural laboratories. A simple experimental demonstration is presented which utilizes part of the hardware and software in the small-scale structural laboratory. Finally, experimental layouts of future potential development and application are addressed and discussed, providing the practitioners with valuable reference for experimental earthquake engineering.

• Journal of Korean Institute of Industrial Engineers
• /
• v.21 no.3
• /
• pp.409-428
• /
• 1995

Job shop scheduling with m different machines and n different jobs is a NP-hard problem of combinatorial optimization. The purpose of the paper is to develop the heuristic method using tabu search for job shop scheduling to minimize makespan or mean flowtime. To apply tabu search to job shop scheduling problem, in this paper we propose the several move methods that employ insert moves in order to generate the neighbor solutions, and present the efficient rescheduling procedure that yields active schedule for a changed operation sequence by a move of operations. We also discuss the tabu search techniques of diversifying the search of solution space as well as the simple tabu search. By experiments, we find the appropriate tabu list size and tabu attributes, and analyze the proposed tabu search techniques with respect to the quality of solutions and the efforts of computation. The experimental results show that the proposed tabu search techniques using long-term memory function have the ability to search a good solution, and are more efficient in the mean flowtime minimization problem than in the makespan minimization.

• Journal of the Korean Crystal Growth and Crystal Technology
• /
• v.4 no.1
• /
• pp.1-10
• /
• 1994

It is well known that crystal growth is an essential area of research for optoelectronic materials. In the present paper the growth techniques and the advanced progress in development of crystal materials important for optoelectronic applications are described. New growth techniques, developed in the authors' laboratory, enabling the introduction of expanded applications of new materials are presented. After a review of recent developments of new techniques for optoelectronic materials, own experimental studies will be discussed.

• International Journal of Automotive Technology
• /
• v.8 no.4
• /
• pp.513-520
• /
• 2007

In this paper, we propose an approach to identify the driving environment for intelligent highway vehicles by means of image processing and computer vision techniques. The proposed approach mainly consists of two consecutive computational steps. The first step is the lane marking detection, which is used to identify the location of the host vehicle and road geometry. In this step, related standard image processing techniques are adapted for lane-related information. In the second step, by using the output from the first step, a four-stage algorithm for vehicle detection is proposed to provide information on the relative position and speed between the host vehicle and each preceding vehicle. The proposed approach has been validated in several real-world scenarios. Herein, experimental results indicate low false alarm and low false dismissal and have demonstrated the robustness of the proposed detection approach.

• Proceedings of the KSRS Conference
• /
• 2004.10a
• /
• pp.721-725
• /
• 2004

Multiple-count problem is occurred when rectangle objects span across several buckets. The Cumulative Density (CD) histogram is a technique which solves multiple-count problem by keeping four sub-histograms corresponding to the four points of rectangle. Although it provides exact results with constant response time, there is still a considerable issue. Since it is based on a query window which aligns with a given grid, a number of errors may be occurred when it is applied to real applications. In this paper, we proposed selectivity estimation techniques using the generalized cumulative density histogram based on two probabilistic models: (1) probabilistic model which considers the query window area ratio, (2) probabilistic model which considers intersection area between a given grid and objects. In order to evaluate the proposed methods, we experimented with real dataset and experimental results showed that the proposed technique was superior to the existing selectivity estimation techniques. The proposed techniques can be used to accurately quantify the selectivity of the spatial range query on rectangle objects.

• Proceedings of the KSME Conference
• /
• 2001.11a
• /
• pp.596-602
• /
• 2001

Structures in current use are required of weight reduction and strength in many instances. This naturally necessitates frequent applications of composite materials in many areas. Elastic constants are one of key parameters in determining design guidelines for the specific applications of particular materials. In this research two vibratory techniques (acoustic resonance method and impulse technique)are utilized to evaluate elastic constants. Both techniques are suitable for the measurements of dynamic elastic constants. The Impulse technique provides a quick method for the measurement while the acoustic resonance method produces the values of elastic constants which agree better with theoretical values.

• Korean Journal of Air-Conditioning and Refrigeration Engineering
• /
• v.16 no.4
• /
• pp.349-354
• /
• 2004

Scale is formed when hard water is heated or cooled in heat transfer equipments such as heat exchangers, condensers, evaporators, cooling towers, boilers, and pipe walls. When the scale deposits in a heat exchanger surface, it is conventionally called fouling. The objective of the present study is to analyze the process of the fouling formation in a heat exchanger according to different types of water using visualization techniques. In order to experimentally investigate the formation of the fouling, this study built a lab-scaled heat exchanging system. Using the visualization techniques of Scanning Electron Microscopy (SEM) and X-Ray diffraction method, the three dimensional configurations of the fouling formation could be successfully obtained. Based on the experimental results, it was found that the configurations of the fouling formation were different when using tap water compared to river water.

• Proceedings of the Korea Society of Environmental Biology Conference
• /
• 1994.05a
• /
• pp.15-36
• /
• 1994

Biological techniques by aquatic macroanimals to improve water quality may be classified overally as three types. The fist is to use aquatic palnts for nutrient removal, the second is to use grass fish to control algae or aquatic plant, and the third is to modify trophostructure by using top-consumer to control algae(top-down effect). This paper includes a general review on above-mentioned techniques, some experimental or field-based results, and some discussion on Korean status and prospect on this part.

• International Journal of Advanced Culture Technology
• /
• v.5 no.3
• /
• pp.40-45
• /
• 2017

This data mining technique was used to extract useful information from percutaneous coronary intervention data obtained from the US public data homepage. The experiment was performed by extracting data on the area, frequency of operation, and the number of deaths. It led us to finding of meaningful correlations, patterns, and trends using various algorithms, pattern techniques, and statistical techniques. In this paper, information is obtained through efficient decision tree and cluster analysis in predicting the incidence of percutaneous coronary intervention and mortality. In the cluster analysis, EM algorithm was used to evaluate the suitability of the algorithm for each situation based on performance tests and verification of results. In the cluster analysis, the experimental data were classified using the EM algorithm, and we evaluated which models are more effective in comparing functions. Using data mining technique, it was identified which areas had effective treatment techniques and which areas were vulnerable, and we can predict the frequency and mortality of percutaneous coronary intervention for heart disease.