• The Journal of the Acoustical Society of Korea
• /
• v.19 no.5
• /
• pp.13-19
• /
• 2000

The blind segmentation method, which segments input speech data into recognition unit without any prior knowledge, plays an important role in continuous speech recognition system and corpus generation. As no prior knowledge is required, this method is rather simple to implement, but in general, it suffers from bad performance when compared to the knowledge-based segmentation method. In this paper, we introduce a method to improve the performance of a blind segmentation of Korean continuous speech by postprocessing the segment boundaries obtained from the blind segmentation. In the preprocessing stage, the candidate boundaries are extracted by a clustering technique based on the GLR(generalized likelihood ratio) distance measure. In the postprocessing stage, the final phoneme boundaries are selected from the candidates by utilizing a simple a priori knowledge on the syllabic structure of Korean, i.e., the maximum number of phonemes between any consecutive nuclei is limited. The experimental result was rather promising : the proposed method yields 25% reduction of insertion error rate compared that of the blind segmentation alone.

• KIISE Transactions on Computing Practices
• /
• v.21 no.1
• /
• pp.94-99
• /
• 2015

With the recent machine learning paradigm of using nonparametric Bayesian statistics and statistical inference based on random sampling, the Dirichlet distribution finds many uses in a variety of graphical models. It is a multivariate generalization of the gamma distribution and is defined on a continuous (K-1)-simplex. This paper presents a sampling method for a Dirichlet distribution for the problem of dividing an integer X into a sequence of K integers which sum to X. The target samples in our problem are all positive integer vectors when multiplied by a given X. They must be sampled from the correspondingly gridded simplex. In this paper we develop a Markov Chain Monte Carlo (MCMC) proposal distribution for the neighborhood grid points on the simplex and then present the complete algorithm based on the Metropolis-Hastings algorithm. The proposed algorithm can be used for the Markov model, HMM, and Semi-Markov model for accurate state-duration modeling. It can also be used for the Gamma-Dirichlet HMM to model q the global-local duration distributions.

• KIPS Transactions on Software and Data Engineering
• /
• v.7 no.7
• /
• pp.259-266
• /
• 2018

In this paper, we propose a method to classify a document using a Recurrent Neural Network by extracting features considering word sense and contexts. Word2vec method is adopted to include the order and meaning of the words expressing the word in the document as a vector. Doc2vec is applied for considering the context to extract the feature of the document. RNN classifier, which includes the output of the previous node as the input of the next node, is used as the document classification method. RNN classifier presents good performance for document classification because it is suitable for sequence data among neural network classifiers. We applied GRU (Gated Recurrent Unit) model which solves the vanishing gradient problem of RNN. It also reduces computation speed. We used one Hangul document set and two English document sets for the experiments and GRU based document classifier improves performance by about 3.5% compared to CNN based document classifier.

• The Journal of the Korea Contents Association
• /
• v.10 no.12
• /
• pp.27-33
• /
• 2010

In this study, we conducted the classification experiments with GMM (Gaussian Mixture Model) from combining the extracted features by using microphone, Gyro sensor and Acceleration sensor in 9 different environment types. Existing studies of Context Aware wanted to recognize the Environment situation mainly using the Environment sound data with microphone, but there was limitation of reflecting recognition owing to structural characteristics of Environment sound which are composed of various noises combination. Hence we proposed the additional application methods which added Gyro sensor and Acceleration sensor data in order to reflect recognition agent's movement feature. According to the experimental results, the method combining Acceleration sensor data with the data of existing Environment sound feature improves the recognition performance by more than 5%, when compared with existing methods of getting only Environment sound feature data from the Microphone.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.43 no.2 s.308
• /
• pp.63-71
• /
• 2006

Denoising and reconstruction of color images are extensively studied in the field of computer vision and image processing. Especially, denoising and reconstruction of color face images are more difficult than those of natural images because of the structural characteristics of human faces as well as the subtleties of color interactions. In this paper, we propose a denoising method based on PCA reconstruction for removing complex color noise on human faces, which is not easy to remove by using vectorial color filters. The proposed method is composed of the following five steps: training of canonical eigenface space using PCA, automatic extraction of facial features using active appearance model, relishing of reconstructed color image using bilateral filter, extraction of noise regions using the variance of training data, and reconstruction using partial information of input images (except the noise regions) and blending of the reconstructed image with the original image. Experimental results show that the proposed denoising method maintains the structural characteristics of input faces, while efficiently removing complex color noise.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.44 no.5
• /
• pp.35-44
• /
• 2007

Since the omnidirectional camera system with a very large field of view could take many information about environment scene from few images, various researches for calibration and 3D reconstruction using omnidirectional image have been presented actively. Most of line segments of man-made objects we projected to the contours by using the omnidirectional camera model. Therefore, the corresponding contours among images sequences would be useful for computing the camera transformations including rotation and translation. This paper presents a novel two step minimization method to estimate the extrinsic parameters of the camera from the corresponding contours. In the first step, coarse camera parameters are estimated by minimizing an angular error function between epipolar planes and back-projected vectors from each corresponding point. Then we can compute the final parameters minimizing a distance error of the projected contours and the actual contours. Simulation results on the synthetic and real images demonstrated that our algorithm can achieve precise contour matching and camera motion estimation.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.22 no.5
• /
• pp.656-662
• /
• 2012

In this paper, we describe a new summarizing method based on a graph-based and a sense-based analysis. In the graph-based analysis, we convert sentences in a document into word vectors and calculate the similarity between each sentence using LSA. We reflect this similarity of sentences and the rarity scores of words in sentences to define weights of edges in the graph. Meanwhile, in the sense-based analysis, in order to determine the sense of words, subjectivity or objectivity, we built a database which is extended from the golden standards using Wordnet. We calculate the subjectivity of sentences from the sense of words, and select more subjective sentences. Lastly, we combine the results of these two methods. We evaluate the performance of the proposed method using classification games, which are usually used to measure the performances of summarization methods. We compare our method with the MS-Word auto-summarization, and verify the effectiveness of ours.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.24 no.6
• /
• pp.665-670
• /
• 2014

A SVM is a kind of binary classifier in order to find optimal hyperplane which separates training data into two groups. Due to its remarkable performance, the SVM is applied in various fields such as inductive inference, binary classification or making predictions. Also it is a representative black box model; there are plenty of actively discussed researches about analyzing trained SVM classifier. This paper conducts a study on a method that is automatically detecting the line-shaped echoes, sun strobe echo and radial interference echo, using the SVM algorithm because the line-shaped echoes appear relatively often and disturb weather forecasting process. Using a spatial clustering method and corrected reflectivity data in the weather radar, the training data is made up with mean reflectivity, size, appearance, centroid altitude and so forth. With actual occurrence cases of the line-shaped echoes, the trained SVM classifier is verified, and analyzed its characteristics using the decision tree method.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.25 no.8B
• /
• pp.1424-1435
• /
• 2000

A bit-rate controller that can achieve a constant bit rate when coding object-based video sequences is an important part to achieve an adaptation to bit-rate constraints, desired video quality, distribution of bits among objects, relationship between texture and shape coding, and determination of frame skip or not. Therefore we design content-based bit rate controller which will be used for relevant bit-rate control. The implementation is an extension of MPEG-4 rate control algorithm which employs a quadratic rate-quantizer model. The importance of different objects in a video is analyzed and segmented into a number of VOPs which are adaptively bit-allocated using the object-based modelling. Some test sequences are observed by a number of non-experts and interests in each object are analysed. The initial total target bit-rate for all objects is obtained by using the proposed technique. Then the total target bits are jointly analyzed for preventing from overflow or underflow of the buffer fullness. The target bits are distributed to each object in view of its importance, not only of statistical analysis such as motion vector magnitude, size of object shape, and coding distortion of previous frame. The scheme is compared with the rate controller adopted by the MPEG-4 VM8 video coder by representing their statistics and performance.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.37 no.1C
• /
• pp.45-53
• /
• 2012

In this paper, we proposed the efficient addressing scheme for improving the performance of routing algorithm by using ZigBee in Smart Grid environment. A distributed address allocation scheme used an existing algorithm that has wasted in address space. Therefore proposing x, y and z coordinate axes from divided address space of 16 bit to solve this problems. Each node was reduced not only bitwise but also multi hop using the coordinate axes while routing than $Cskip$ algorithm. I compared the performance between the standard and the proposed mechanism through the numerical analysis. Simulation verified performance about decrease averaging multi hop count that compare proposing algorithm and another. The numerical analysis results show that proposed algorithm reduced the multi hop better than ZigBee distributed address assignmen.