• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.20 no.2
• /
• pp.401-407
• /
• 2016

We can consider the following problems for two given points p and q in a simple polygon P. (1) Compute the set of points of P which are visible from both p and q. (2) Compute the set of points of P which are visible from either p or q. They are corresponding to the problems which are to compute the intersection and the union of two visibility polygons. In this paper, we consider algorithms for solving these problems on a reconfigurable mesh(in short, RMESH). The algorithm in [1] can compute the intersection of two general polygons in constant time on an RMESH with size O($n^3$), where n is the total number of vertices of two polygons. In this paper, we construct the planar subdivision graph in constant time on an RMESH with size O($n^2$) using the properties of the visibility polygon for preprocessing. Then we present O($log^2n$) time algorithms for computing the union as well as the intersection of two visibility polygons, which improve the processor-time product from O($n^3$) to O($n^2log^2n$).

• The Journal of the Korea Contents Association
• /
• v.11 no.12
• /
• pp.497-503
• /
• 2011

In this paper, we proposed the MCE as a GMM training method to improve the performance of environmental sounds recognition. We model the environmental sounds data with newly defined misclassification function using the log likelihood of the corresponding class and the log likelihood of the rest classes for discriminative training. The model parameters are estimated with the loss function using GPD(generalized probabilistic descent). For recognition performance comparison, we extracted the 12 degrees features using preprocessing and MFCC(mel-frequency cepstral coefficients) of the 9 kinds of environmental sounds and carry out GMM classification experiments. According to the experimental results, MCE training method showed the best performance by an average of 87.06% with 19 mixtures. This result confirmed us that MCE training method could be effectively used as a GMM training method in environmental sounds recognition.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.10 no.3
• /
• pp.608-613
• /
• 2006

There are many e-showing mall because of revitalization of e-commerce system. It is necessary to recommending system of products that is for saving time and effort of customer. In this paper, we propose the system that is applying classification among data mining techniques to analysis of log data of customer. This log data contains access of user and purchasing of products. The proposed system operates in two phases. The first phase is composed of data filter module and association extraction module among web pages. The second phase is composed of personalization module and rule generation module. Customer can easily know the recommended sites because the proposed system can present rank of the recommended web pages to customer. As a result, the proposed system can efficiently do recommending of products to customer.

• Geophysics and Geophysical Exploration
• /
• v.1 no.3
• /
• pp.183-187
• /
• 1998

Seismic waveform estimation is based on the assumption that the seismic trace tying a well is one dimensional convolution of the propagating seismic waveform and the reflectivity series derived from well logs (sonic and density). With this assumption, the waveform embedded in a seismic trace can be estimated using a Wiener match filter. In this paper, I experimented a preprocessing procedure that applies both on the seismic trace and on the reflectivity series. The procedure is based on the assumption that the travel time can be estimated better from the seismic trace and that the instantaneous reflectivity values can be measured better on the well log. Thus the procedure is, 1) start-time adjustment and dynamic differential stretches are applied on the sonic log, and 2) seismic amplitudes are balanced such that the low frequency part of the seismic are matched to that of the reflectivities derived from well logs.

• International conference on construction engineering and project management
• /
• 2020.12a
• /
• pp.463-481
• /
• 2020

Cutter suction dredgers (CSDs) are widely used in various dredging constructions such as channel excavation, wharf construction, and reef construction. During a CSD construction, the main operation is to control the swing speed of cutter to keep the slurry concentration in a proper range. However, the slurry concentration cannot be monitored in real-time, i.e., there is a "time-lag effect" in the log of slurry concentration, making it difficult for operators to make the optimal decision on controlling. Concerning this issue, a solution scheme that using real-time monitored indicators to predict current slurry concentration is proposed in this research. The characteristics of the CSD monitoring data are first studied, and a set of preprocessing methods are presented. Then we put forward the concept of "index class" to select the important indices. Finally, an ensemble learning algorithm is set up to fit the relationship between the slurry concentration and the indices of the index classes. In the experiment, log data over seven days of a practical dredging construction is collected. For comparison, the Deep Neural Network (DNN), Long Short Time Memory (LSTM), Support Vector Machine (SVM), Random Forest (RF), Gradient Boosting Decision Tree (GBDT), and the Bayesian Ridge algorithm are tried. The results show that our method has the best performance with an R² of 0.886 and a mean square error (MSE) of 5.538. This research provides an effective way for real-time predicting the slurry concentration of CSDs and can help to improve the stationarity and production efficiency of dredging construction.

• Journal of KIISE
• /
• v.45 no.3
• /
• pp.294-302
• /
• 2018

With the emergence of the new service industry due to the development of information and communication technology, cyber space risks such as personal information infringement and industrial confidentiality leakage have diversified, and the security problem has emerged as a critical issue. In this paper, we propose a behavior-based anomaly detection method that is suitable for real-time and large-volume data analysis technology. We show that the proposed detection method is superior to existing signature security countermeasures that are based on large-capacity user log data according to in-company personal information abuse and internal information leakage. As the proposed behavior-based anomaly detection method requires a technique for processing large amounts of data, a real-time search engine is used, called Elasticsearch, which is based on an inverted index. In addition, statistical based frequency analysis and preprocessing were performed for data analysis, and the DBSCAN algorithm, which is a density based clustering method, was applied to classify abnormal data with an example for easy analysis through visualization. Unlike the existing anomaly detection system, the proposed behavior-based anomaly detection technique is promising as it enables anomaly detection analysis without the need to set the threshold value separately, and was proposed from a statistical perspective.

• Proceedings of the Korean Society of Near Infrared Spectroscopy Conference
• /
• 2001.06a
• /
• pp.1530-1530
• /
• 2001

The objectives of this study were to develop models to predict quality parameters of Korean bee-honeys by visible and NIR spectroscopic technique. Two kinds of bee-honey fronl acacia and polyflower sources were tested in this study. The honeys were harvested in the spring of 2000 and stored in the storage facility at 20$^{\circ}C$ during experiments. Total of 394 samples of honey were analyzed. Reflectance spectra, moisture contents, ash, invert sugar, sucrose, F/G (fructose/glucose) ratio, HMF (hydroxymethyl furfural), and C12/C13 ratio of honeys were measured. The average values for the tested honeys were 19.9% of moisture contents, 0.12% of ash, 68.4% of invert sugar, 5.7% of sucrose, 1.27 of F/G(fructose/glucose) ratio, 14.4 mg/kg of HMF, and -19.1 of C12/C13 ratio. A spectrophotometer, equipped with a single-beam scanning monochromator (NIR Systems, Model 6500, USA) and a horizontal setup module, was used to collect reflectance data from honey. The reflectance spectra were measured in wavelength ranges of 400∼2,498 nm. with 2 nm of interval. Thirty-two repetitive scans were averaged, transformed to log(1/Reflectance), and then were stored in a microcomputer file, forming one spectrum per measurement. A sample cell and reflectance plate were made to hold honey samples constantly. Spectra of honey samples were divided into a calibration set and a validation set. The calibration set was used during model development, and the validation set was used to predict quality parameters from unknown spectra. The PLS(Partial Least Square) models were developed to predict the quality parameters of honeys. The first and the second derivatives of raw spectra were also used to develop the models with proper smoothing gap. The MSC (multiplicative scatter correction) and the SNV & Dtr.(standard normal variate and detranding) preprocessing were applied to all spectra to minimize sample-to-sample light scatter differences. The PLS models showed good relationships between predicted and measured quality parameters of honeys in the wavelength range of 1100∼2200 nm. However, the PLS analysis was not good enough to predict HMF of honeys.