• The Journal of the Acoustical Society of Korea
• /
• v.24 no.3
• /
• pp.150-159
• /
• 2005

This paper addresses a voice personality transformation algorithm which makes one person's voices sound as if another person's voices. In the proposed method, one person's voices are represented by LPC cepstrum, pitch period and speaking rate, the appropriate transformation rules for each Parameter are constructed. The Gaussian Mixture Model (GMM) is used to model one speaker's LPC cepstrums and conditional probability is used to model the relationship between two speaker's LPC cepstrums. To obtain the parameters representing each probabilistic model. a Maximum Likelihood (ML) estimation method is employed. The transformed LPC cepstrums are obtained by using a Minimum Mean Square Error (MMSE) criterion. Pitch period and speaking rate are used as the parameters for prosody transformation, which is implemented by using the ratio of the average values. The proposed method reveals the superior performance to the previous VQ-based method in subjective measures including average cepstrum distance reduction ratio and likelihood increasing ratio. In subjective test. we obtained almost the same correct identification ratio as the previous method and we also confirmed that high qualify transformed speech is obtained, which is due to the smoothly evolving spectral contours over time.

• Journal of the Institute of Convergence Signal Processing
• /
• v.10 no.2
• /
• pp.120-126
• /
• 2009

There are many different types of surface defects on semiconductor Integrated Chips (IC's) caused by various factors during manufacturing process, such as cracks, foreign materials, chip-outs, chips, and voids. These defects must be detected and classified by an inspection system for productivity improvement and effective process control. Among defects, in particular, foreign materials and chips are the most difficult ones to classify accurately. A vision system composed of a carefully designed optical system and a processing algorithm is proposed to detect and classify the defects on QFN(Quad Flat No-leads) packages. The processing algorithm uses features derived from the defect's position and brightness value in the Maximum Likelihood classifier and the optical system is designed to effectively extract the features used in the classifier. In experiments we confirm that this method gives more effective result in classifying foreign materials and chips.

• Progress in Medical Physics
• /
• v.23 no.4
• /
• pp.309-316
• /
• 2012

Multimodal-imaging technique has been rapidly developed for improvement of diagnosis and evaluation of therapeutic effects. In despite of integrated hardware, registration accuracy was decreased due to a discrepancy between multimodal image and insufficiency of count in accordance with different acquisition method of each modality. The purpose of this study was to improve the PET image by event data resampling through analysis of data format, noise and statistical properties of small animal PET list data. Inveon PET listmode data was acquired as static data for 10 min after 60 min of 37 MBq/0.1 ml $^{18}F$-FDG injection via tail vein. Listmode data format was consist of packet containing 48 bit in which divided 8 bit header and 40 bit payload space. Realigned sinogram was generated from resampled event data of original listmode by using adjustment of LOR location, simple event magnification and nonparametric bootstrap. Sinogram was reconstructed for imaging using OSEM 2D algorithm with 16 subset and 4 iterations. Prompt coincidence was 13,940,707 count measured from PET data header and 13,936,687 count measured from analysis of list-event data. In simple event magnification of PET data, maximum was improved from 1.336 to 1.743, but noise was also increased. Resampling efficiency of PET data was assessed from de-noised and improved image by shift operation of payload value of sequential packet. Bootstrap resampling technique provides the PET image which noise and statistical properties was improved. List-event data resampling method would be aid to improve registration accuracy and early diagnosis efficiency.

• Journal of Korea Multimedia Society
• /
• v.5 no.5
• /
• pp.512-522
• /
• 2002

In this paper, we propose a hybrid method which is mixed with frame selection and weighting model rank method, based on GMM(gaussian mixture model), for real-time text-independent speaker recognition system. In the system, maximum likelihood estimation was used for GMM parameter optimization, and maximum likelihood was used for recognition basically Proposed hybrid method has two steps. First, likelihood score was calculated with speaker models and test data at frame level, and the difference is calculated between the biggest likelihood value and second. And then, the frame is selected if the difference is bigger than threshold. The second, instead of calculated likelihood, weighting value is used for calculating total score at each selected frame. Cepstrum coefficient and regressive coefficient were used as feature parameters, and the database for test and training consists of several data which are collected at different time, and data for experience are selected randomly In experiments, we applied each method to baseline system, and tested. In speaker recognition experiments, proposed hybrid method has an average of 4% higher recognition accuracy than frame selection method and 1% higher than W method, implying the effectiveness of it.

• Convergence Security Journal
• /
• v.21 no.2
• /
• pp.11-18
• /
• 2021

To protect tangible and intangible assets, most of the companies are conducting information protection monitoring by using various security equipment in the IT service network. As the security equipment that needs to be protected increases in the process of upgrading and expanding the service network, it is difficult to monitor the possible exposure to the attack for the entire service network. As a countermeasure to this, various studies have been conducted to detect external attacks and illegal communication of equipment, but studies on effective monitoring of the open service ports and construction of illegal communication monitoring system for large-scale service networks are insufficient. In this study, we propose a framework that can monitor information leakage and illegal communication attempts in a wide range of service networks without large-scale investment by analyzing 'Netflow statistical information' of backbone network equipment, which is the gateway to the entire data flow of the IT service network. By using machine learning algorithms to the Netfllow data, we could obtain the high classification accuracy of 94% in identifying whether the Telnet service port of operating equipment is open or not, and we could track the illegal communication of the damaged equipment by using the illegal communication history of the damaged equipment.

• Journal of Internet of Things and Convergence
• /
• v.10 no.1
• /
• pp.39-45
• /
• 2024

Container volume is a very important factor in accurate evaluation of port performance, and accurate prediction of effective port development and operation strategies is essential. However, it is difficult to improve the accuracy of container volume prediction due to rapid changes in the marine industry. To solve this problem, it is necessary to analyze the impact on port performance using the Internet of Things (IoT) and apply it to improve the competitiveness and efficiency of Busan Port. Therefore, this study aims to develop a prediction model for predicting the future container volume of Busan Port, and through this, focuses on improving port productivity and making improved decision-making by port management agencies. In order to predict port container volume, this study introduced the Extreme Gradient Boosting (XGBoost) technique of a machine learning model. XGBoost stands out of its higher accuracy, faster learning and prediction than other algorithms, preventing overfitting, along with providing Feature Importance. Especially, XGBoost can be used directly for regression predictive modelling, which helps improve the accuracy of the volume prediction model presented in previous studies. Through this, this study can accurately and reliably predict container volume by the proposed method with a 4.3% MAPE (Mean absolute percentage error) value, highlighting its high forecasting accuracy. It is believed that the accuracy of Busan container volume can be increased through the methodology presented in this study.

• The Journal of the Acoustical Society of Korea
• /
• v.20 no.4
• /
• pp.35-41
• /
• 2001

This paper describes Linear Discriminant Analysis and common vector extraction for speech recognition. Voice signal contains psychological and physiological properties of the speaker as well as dialect differences, acoustical environment effects, and phase differences. For these reasons, the same word spelled out by different speakers can be very different heard. This property of speech signal make it very difficult to extract common properties in the same speech class (word or phoneme). Linear algebra method like BT (Karhunen-Loeve Transformation) is generally used for common properties extraction In the speech signals, but common vector extraction which is suggested by M. Bilginer et at. is used in this paper. The method of M. Bilginer et al. extracts the optimized common vector from the speech signals used for training. And it has 100％ recognition accuracy in the trained data which is used for common vector extraction. In spite of these characteristics, the method has some drawback-we cannot use numbers of speech signal for training and the discriminant information among common vectors is not defined. This paper suggests advanced method which can reduce error rate by maximizing the discriminant information among common vectors. And novel method to normalize the size of common vector also added. The result shows improved performance of algorithm and better recognition accuracy of 2％ than conventional method.

• Korean Chemical Engineering Research
• /
• v.61 no.4
• /
• pp.523-541
• /
• 2023

As accessibility to 3D printers increases, there is a growing frequency of exposure to chemicals associated with 3D printing. However, research on the toxicity and harmfulness of chemicals generated by 3D printing is insufficient, and the performance of toxicity prediction using in silico techniques is limited due to missing molecular structure data. In this study, quantitative structure-activity relationship (QSAR) model based on data-centric AI approach was developed to predict the toxicity of new 3D printing materials by imputing missing values in molecular descriptors. First, MissForest algorithm was utilized to impute missing values in molecular descriptors of hazardous 3D printing materials. Then, based on four different machine learning models (decision tree, random forest, XGBoost, SVM), a machine learning (ML)-based QSAR model was developed to predict the bioconcentration factor (Log BCF), octanol-air partition coefficient (Log Koa), and partition coefficient (Log P). Furthermore, the reliability of the data-centric QSAR model was validated through the Tree-SHAP (SHapley Additive exPlanations) method, which is one of explainable artificial intelligence (XAI) techniques. The proposed imputation method based on the MissForest enlarged approximately 2.5 times more molecular structure data compared to the existing data. Based on the imputed dataset of molecular descriptor, the developed data-centric QSAR model achieved approximately 73%, 76% and 92% of prediction performance for Log BCF, Log Koa, and Log P, respectively. Lastly, Tree-SHAP analysis demonstrated that the data-centric-based QSAR model achieved high prediction performance for toxicity information by identifying key molecular descriptors highly correlated with toxicity indices. Therefore, the proposed QSAR model based on the data-centric XAI approach can be extended to predict the toxicity of potential pollutants in emerging printing chemicals, chemical process, semiconductor or display process.