• Journal of the Korea Society of Computer and Information
• /
• v.14 no.7
• /
• pp.17-24
• /
• 2009

Increasing use of information system has led to larger amount of developing expenses and demands on software. Until recent days, the model using regression analysis based on statistical algorithm has been used. However, Machine learning is more investigated now. This paper estimates the software cost using SVR(Support Vector Regression). a sort of machine learning technique. Also, it finds the best set of parameters applying immune algorithm. In this paper, software cost estimation is performed by SVR based on immune algorithm while changing populations, memory cells, and number of allele. Finally, this paper analyzes and compares the result with existing other machine learning methods.

• Journal of Korea Spatial Information System Society
• /
• v.5 no.1 s.9
• /
• pp.39-47
• /
• 2003

The purpose of this paper is to present the method of acquiring 3D GIS data using the statistical estimating methods of Well Log for balancing Seismic data. We use the reflection coefficients of seismic data to get the parameters for the reservoir characterization and we balance the reflection coefficients of seismic data using well log to increase the confidence of the estimated result. Well logs are required to balance the reflection coefficients at the point where seismic data are acquired. In this research, we discuss the geostatistical estimation methods and we applied these methods to real data. Kriging gives high weights to the close well logs, which means estimated results are mainly affected by close well log. High value of cross variograms gave big difference on cokriging result comparing to kriging results and low value of cross variogram gave little differences.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.13 no.8
• /
• pp.1513-1526
• /
• 2009

The SVC (Scalable Video Coding) scheme can be effectively used for reducing the redundancy and for improving the coding efficiency but, it requires very high computational complexities. In order to accelerate the successful standardization and commercialization of the Advanced Terrestrial-DMB service, it is necessary to overcome this problem. For this aim, in this paper, we propose an efficient aggregated encoder control algorithm, which shows better performances than the conventional control scheme. Computer simulation result shows that the proposed scheme performs about up to 0.3dB better than those of the conventional scheme. Additionally, based on this control scheme, we propose a fast mode decision method by constraining the redundant coding modes based on the statistical properties of the quantization parameter in the spatial scalable encoder. Through computer simulations, it is shown that the proposed control schemes reduce the heavy computational burden up to 12% compared to the conventional scheme, while keeping the objective visual qualify very high.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.31 no.2A
• /
• pp.110-118
• /
• 2006

In part-I of the paper, two sub-channel structures, either contiguous or distributed, were considered. Modeling the SNR distribution over a sub-channel as Ricean in general, the statistical chracteristics were investigated. In this part of the paper, we develop a generalized two step channel/resource allocation algorithm, which incorporates the two statistical measurements, and analyze the spectral efficiency of OFDMA in terms of average frequency utilization for the two sub-channel structures. In OFDMA with distributed structure, the key design parameter would be the sub-channel bandwidth. To give an insight into the impact on this parameter, we show in the numerical results the frequency utilization as a function of sub-channel bandwidth normalized to coherence bandwidth. As confirmed by numerical results, for contiguous sub-channel structure, we obtain the nominal multiuser diversity gain when the sub-channel bandwidth is smaller than the coherence bandwidth and lose the gain as it is getting larger.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.31 no.2A
• /
• pp.119-127
• /
• 2006

In this two-part paper, we consider dynamic resource allocation in orthogonal frequency division multiple access(OFDMA). To reduce the reverse link overhead for channel quality information(CQI) feedback, a set of sub-carriers are tied up to a sub-channel to be used as the unit of CQI feedback, user-multiplexing and the corresponding power/rate allocation. Specifically, we focus on two sub-channel structures, either aggregated or distributed, where the SNR distribution over a sub-channel is modeled as Ricean in general, and the channel quality of a sub-channel is summarized as the mean and variance of channel gain envelop divided by noise standard deviation. Then, we develop a generalized two step channel/resource allocation algorithm, which uses the two statistical measurements, and analyze the spectral efficiency of the OFDMA system in terms of average frequency utilization. An extension to proportional fair algorithm will also be addressed. As confirmed by numerical results, the aggregated structure is preferred especially when intending aggressive link adaptation.

• Journal of the Korea Computer Graphics Society
• /
• v.28 no.4
• /
• pp.1-11
• /
• 2022

The Skinned Multi-Person Linear Model (SMPL) is the most widely used parametric 3D Human Model optimized and learned from CAESAR, a 3D human scanned database created with measurements from 3,800 people living in United States in the 1990s. We point out the lack of racial diversity of body types in SMPL and propose K-SMPL that better represents Korean 3D body shapes. To this end, we develop a fitting algorithm to estimate 2,773 Korean 3D body shapes from Korean body measurement data. By conducting principle component analysis to the estimated Korean body shapes, we construct K-SMPL model that can generate various Korean body shape in 3D. K-SMPL model allows to improve the fitting accuracy over SMPL with respect to the Korean body measurement data. K-SMPL model can be widely used for avatar generation and human shape fitting for Korean.

• The Journal of the Convergence on Culture Technology
• /
• v.9 no.5
• /
• pp.569-582
• /
• 2023

Lead-acid Battery is the oldest rechargeable battery system and has maintained its position in the rechargeable battery field. The battery causes thermal runaway for various reasons, which can lead to major accidents. Therefore, preventing thermal runaway is a key part of the battery management system. Recently, research is underway to categorize thermal runaway battery cells into machine learning. In this paper, we present a thermal runaway hazard cell detection and verification algorithm using DBSCAN and statistical method. An experiment was conducted to classify thermal runaway hazard cells using only the resistance values as measured by the Battery Management System (BMS). The results demonstrated the efficacy of the proposed algorithms in accurately classifying thermal runaway cells. Furthermore, the proposed algorithm was able to classify thermal runaway cells between thermal runaway hazard cells and cells containing noise. Additionally, the thermal runaway hazard cells were early detected through the optimization of DBSCAN parameters using a grid search approach.

• Journal of the Korea Computer Graphics Society
• /
• v.17 no.1
• /
• pp.9-16
• /
• 2011

In this paper, we describe the development of a system for generating a 3-dimensional human face and predicting it's appearance as it ages over subsequent years using 3D scanned facial data and photo images. It is composed of 3-dimensional texture mapping functions, a facial definition parameter input tool, and 3-dimensional facial prediction algorithms. With the texture mapping functions, we can generate a new model of a given face at a specified age using a scanned facial model and photo images. The texture mapping is done using three photo images - a front and two side images of a face. The facial definition parameter input tool is a user interface necessary for texture mapping and used for matching facial feature points between photo images and a 3D scanned facial model in order to obtain material values in high resolution. We have calculated material values for future facial models and predicted future facial models in high resolution with a statistical analysis using 100 scanned facial models.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.12 no.6
• /
• pp.571-576
• /
• 2002

There have been a lot of considerations dealing with improving the performance of neuro-fuzzy system. The studies on the neuro-fuzzy modeling have largely been devoted to two approaches. First is to improve performance index of system. The other is to reduce the structure size. In spite of its satisfactory result, it should be noted that these are difficult to extend to high dimensional input or to increase the membership functions. We propose a novel neuro-fuzzy system based on the efficient clustering method for initializing the parameters of the premise part. It is a very useful method that maintains a few number of rules and improves the performance. It combine the various algorithms to improve the performance. The Expectation-Maximization algorithm of Gaussian mixture model is an efficient estimation method for unknown parameter estimation of mirture model. The obtained parameters are used for fuzzy clustering method. The proposed method satisfies these two requirements using the Gaussian mixture model and neuro-fuzzy modeling. Experimental results indicate that the proposed method is capable of giving reliable performance.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.7B
• /
• pp.585-597
• /
• 2004

In this paper, we propose the functional models of optical burst switching (OBS) routers to apply GMPLS (Generalized Multiprotocol Label Switching) to optical networks based on OBS. In addition, we introduce a connection admission control (CAC) algorithms which are operated in this models and can accommodate the required QoS. Firstly, the characteristics of current GMPLS and OBS for the optical Internet are basically considered. With this consideration, the models are proposed to accept OBS features which include the recognition of data boundary with control information and the statistical multiplexing in terms of bursts. Secondly, we use an offset time decision (OTD) algorithm on behalf of controlling the connection admission with taking QoS parameters such as burst loss rate(BLR) and service-differentiation ratio(SDR) into consideration. The proposed CAC algorithms use the offered load of LSP (Label Switched Path), wavelength information, and QoS parameters as inputs of OTD algorithm. A call setup request will be accepted when the offset time decided by OTD algorithm is reasonable for guaranteeing its requested QoS. Simulation is used for performance evaluation. Results show the proposed schemes can guarantee the required QoS and those are better than the previous one in terms of channel utilization.