• 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.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.41 no.9
• /
• pp.1120-1131
• /
• 2016

In this paper, we propose a smoker recognition algorithm, detecting smokers in a video sequence in order to prevent fire accidents. We use description-based method in hierarchical approaches to recognize smoker's activity, the algorithm consists of background subtraction, object detection, event search, event judgement. Background subtraction generates slow-motion and fast-motion foreground image from input image using Gaussian mixture model with two different learning-rate. Then, it extracts object locations in the slow-motion image using chain-rule based contour detection. For each object, face is detected by using Haar-like feature and smoke is detected by reflecting frequency and direction of smoke in fast-motion foreground. Hand movements are detected by motion estimation. The algorithm examines the features in a certain interval and infers that whether the object is a smoker. It robustly can detect a smoker among different objects while achieving real-time performance.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.21 no.6
• /
• pp.171-181
• /
• 2011

Proposed in this paper is the intelligent video surveillance system to effectively detect multiple loitering objects even that disappear from the out of camera's field of view and later return to a target zone. After the background and foreground are segmented using Gaussian mixture model and shadows are removed, the objects returning to the target zone is recognized using the chromaticity histogram and the duration of loitering is preserved. For more accurate measurement of the loitering behavior, the camera calibration is also applied to map the image plane to the real-world ground. Hence, the loitering behavior can be detected by considering the time duration of the object's existence in the real-world space. The experiment was performed using loitering video and all of the loitering behaviors are accurately detected.

• Journal of Convergence for Information Technology
• /
• v.12 no.3
• /
• pp.252-258
• /
• 2022

In this study, we propose a method for analyzing students' understanding of probability and statistics in small lectures at universities. A computer-based test for probability and statistics was performed on 95 science and engineering college students. After dividing the students' responses into 7 clusters using the Robust PCA and the Gaussian mixture model, the achievement of each subject was analyzed for each cluster. High-ranking clusters generally showed high achievement on most topics except for statistical estimation, and low-achieving clusters showed strengths and weaknesses on different topics. Compared to the widely used PCA-based dimension reduction followed by clustering analysis, the proposed method showed each group's characteristics more clearly. The characteristics of each cluster can be used to develop an individualized learning strategy.

• Journal of Internet Computing and Services
• /
• v.18 no.4
• /
• pp.61-67
• /
• 2017

Recently, CCTV has been combined with areas such as big data, artificial intelligence, and image analysis to detect various abnormal behaviors and to detect and analyze the overall situation of objects such as people. Image analysis research for this intelligent video surveillance function is progressing actively. However, CCTV images using 2D information generally have limitations such as object misrecognition due to lack of topological information. This problem can be solved by adding the depth information of the object created by using two cameras to the image. In this paper, we perform background modeling using Mixture of Gaussian technique and detect whether there are moving objects by segmenting the foreground from the modeled background. In order to perform the depth information-based segmentation using the RGB information-based segmentation results, stereo-based depth maps are generated using two cameras. Next, the RGB-based segmented region is set as a domain for extracting depth information, and depth-based segmentation is performed within the domain. In order to detect the center point of a robustly segmented object and to track the direction, the movement of the object is tracked by applying the CAMShift technique, which is the most basic object tracking method. From the experiments, we prove the efficiency of the proposed object detection and tracking method using the RGB-D model.

• KIPS Transactions on Software and Data Engineering
• /
• v.6 no.11
• /
• pp.537-542
• /
• 2017

Aggressive driving is a major cause of car accidents. Previous studies have mainly analyzed young driver's aggressive driving tendency, yet they were only done through pure clustering or classification technique of machine learning. However, since elderly people have different driving habits due to their fragile physical conditions, it is necessary to develop a new method such as enhancing the characteristics of driving data to properly analyze aggressive driving of elderly drivers. In this study, acceleration data collected from a smartphone of a driving vehicle is analyzed by a newly proposed ECA(Enhanced Clustering method for Acceleration data) technique, coupled with a conventional clustering technique (K-means Clustering, Expectation-maximization algorithm). ECA selects high-intensity data among the data of the cluster group detected through K-means and EM in all of the subjects' data and models the characteristic data through the scaled value. Using this method, the aggressive driving data of all youth and elderly experiment participants were collected, unlike the pure clustering method. We further found that the K-means clustering has higher detection efficiency than EM method. Also, the results of K-means clustering demonstrate that a young driver has a driving strength 1.29 times higher than that of an elderly driver. In conclusion, the proposed method of our research is able to detect aggressive driving maneuvers from data of the elderly having low operating intensity. The proposed method is able to construct a customized safe driving system for the elderly driver. In the future, it will be possible to detect abnormal driving conditions and to use the collected data for early warning to drivers.

• Clinical and Experimental Pediatrics
• /
• v.48 no.11
• /
• pp.1179-1186
• /
• 2005

Purpose : A universal standard of the birth weight for gestational age cannot be made since girth weight distribution varies with race and other sociodemographic factors. This report aims to establish the birth weight distribution curve by gestational age, specific for Korean live births. Methods : We used the national birth certificate data of all live births in Korea from January 2001 to December 2003; for live births with gestational ages 24 weeks to 44 weeks(n=1,509,763), we obtained mean birth weigh, standard deviation and 10th, 25th, 50th, 75th and 90th percentile values for each gestational age group by one week increment. Then, we investigated the birth weight distribution of each gestational age group by the normal Gaussian model. To establish final standard values of Korean birth weight distribution by gestational age, we used the finite mixture model to eliminate erroneous birth slights for respective gestational ages. Results : For gestational ages 28 weeks 32 weeks, birth weight distribution showed a biologically implausible skewed tail or bimodal distribution. Following correction of the erroneous distribution by using the finite mixture model, the constructed curve of birth weight distribution was compared to those of other studies. The Korean birth weight percentile values were generally lower than those for Norwegians and North Americans, particularly after 37 weeks of gestation. The Korean curve was similar to that of Lubchenco both 50th and 90th percentiles, but generally the Korean curve had higher 10th percentile values. Conclusion : This birth weight distribution curve by gestational age is based on the most recent and the national population data compared to previous studies in Korea. We hope that for Korean infants, this curve will help clinicians in defining and managing the large for gestational age infants and also for infants with intrauterine growth retardation.

• Journal of KIISE:Software and Applications
• /
• v.34 no.1
• /
• pp.1-10
• /
• 2007

This paper proposes a new indexing mechanism for MPEG-7 visual descriptors, especially Dominant Color and Contour Shape descriptors, that guarantees an efficient similarity search for the multimedia database whose visual meta-data are represented with MPEG-7. Since the similarity metric used in the Dominant Color descriptor is based on Gaussian mixture model, the descriptor itself could be transform into a color histogram in which the distribution of the color values follows the Gauss distribution. Then, the transformed Dominant Color descriptor (i.e., the color histogram) is indexed in the proposed indexing mechanism. For the indexing of Contour Shape descriptor, we have used a two-pass algorithm. That is, in the first pass, since the similarity of two shapes could be roughly measured with the global parameters such as eccentricity and circularity used in Contour shape descriptor, the dissimilar image objects could be excluded with these global parameters first. Then, the similarities between the query and remaining image objects are measured with the peak parameters of Contour Shape descriptor. This two-pass approach helps to reduce the computational resources to measure the similarity of image objects using Contour Shape descriptor. This paper also proposes two integration schemes of visual descriptors for an efficient retrieval of multimedia database. The one is to use the weight of descriptor as a yardstick to determine the number of selected similar image objects with respect to that descriptor, and the other is to use the weight as the degree of importance of the descriptor in the global similarity measurement. Experimental results show that the proposed indexing and integration schemes produce a remarkable speed-up comparing to the exact similarity search, although there are some losses in the accuracy because of the approximated computation in indexing. The proposed schemes could be used to build a multimedia database represented in MPEG-7 that guarantees an efficient retrieval.

• The KIPS Transactions:PartB
• /
• v.16B no.2
• /
• pp.131-140
• /
• 2009

This paper presents a new algorithm to the segmentation of the FISH images. First, for segmentation of the cell nuclei from background, a threshold is estimated by using the gaussian mixture model and maximizing the likelihood function of gray value of cell images. After nuclei segmentation, overlapped nuclei and isolated nuclei need to be classified for exact nuclei analysis. For nuclei classification, this paper extracted the morphological features of the nuclei such as compactness, smoothness and moments from training data. Three probability density functions are generated from these features and they are applied to the proposed Bayesian networks as evidences. After nuclei classification, segmenting of overlapped nuclei into isolated nuclei is necessary. This paper first performs intensity gradient transform and watershed algorithm to segment overlapped nuclei. Then proposed stepwise merging strategy is applied to merge several fragments in major nucleus. The experimental results using FISH images show that our system can indeed improve segmentation performance compared to previous researches, since we performed nuclei classification before separating overlapped nuclei.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.12 no.4
• /
• pp.772-780
• /
• 2008

K-means is one of the simplest unsupervised learning algorithms that solve the clustering problem. However K-means suffers the basic shortcoming: the number of clusters k has to be known in advance. In this paper, we propose extensions of X-means, which can estimate the number of clusters using Bayesian information criterion(BIC). We introduce two different versions of algorithm: modified X-means(MX-means) and generalized X-means(GX-means), which employ one full covariance matrix for one cluster and so can estimate the number of clusters efficiently without severe over-fitting which X-means suffers due to its spherical cluster assumption. The algorithms start with one cluster and try to split a cluster iteratively to maximize the BIC score. The former uses K-means algorithm to find a set of optimal clusters with current k, which makes it simple and fast. However it generates wrongly estimated centers when the clusters are overlapped. The latter uses EM algorithm to estimate the parameters and generates more stable clusters even when the clusters are overlapped. Experiments with synthetic data show that the purposed methods can provide a robust estimate of the number of clusters and cluster parameters compared to other existing top-down algorithms.