• The Journal of the Acoustical Society of Korea
• /
• 제19권5호
• /
• pp.3-8
• /
• 2000

In this paper, we studied design and implementation of a robust speech recognition system in noisy car environment. The reference pattern used in the system is DMS(Dynamic Multi-Section). Two separate acoustic models, which are selected automatically depending on the noisy car environment for the speech in a car moving at below 80km/h and over 80km/h are proposed. PLP(Perceptual Linear Predictive) of order 13 is used for the feature vector and OSDP (One-Stage Dynamic Programming) is used for decoding. The system also has the function of editing the phone-book for voice dialing. The system yields a recognition rate of 89.75% for male speakers in SI (speaker independent) mode in a car running on a cemented express way at over 80km/h with a vocabulary of 33 words. The system also yields a recognition rate of 92.29% for male speakers in SI mode in a car running on a paved express way at over 80km/h.

• Journal of Digital Convergence
• /
• 제15권12호
• /
• pp.341-347
• /
• 2017

As the amount of document data increases, the user needs summarized information to understand the document. However, existing document summary research methods rely on overly simple statistics, so there is insufficient research on multiple document summaries for ambiguity of sentences and meaningful sentence generation. In this paper, we investigate semantic connection and preprocessing process to process unnecessary information. Based on the vocabulary semantic pattern information, we propose a multi-document summarization method that enhances semantic connectivity between sentences using VAE. Using sentence word vectors, we reconstruct sentences after learning from compressed information and attribute discriminators generated as latent variables, and semantic connection processing generates a natural summary sentence. Comparing the proposed method with other document summarization methods showed a fine but improved performance, which proved that semantic sentence generation and connectivity can be increased. In the future, we will study how to extend semantic connections by experimenting with various attribute settings.

• Journal of the Korean Institute of Intelligent Systems
• /
• 제26권4호
• /
• pp.259-266
• /
• 2016

Recently, CCTV cameras are emplaced actively to reinforce security and intelligent surveillance systems have been under development for detecting and monitoring of the objects in the video. In this study, we propose a method for detection of upper body in intelligent surveillance system using FCM-based RBFNN classifier realized with the aid of HOG features. Firstly, HOG features that have been originally proposed to detect the pedestrian are adopted to train the unique gradient features about upper body. However, HOG features typically exhibit a very high dimension of which is proportional to the size of the input image, it is necessary to reduce the dimension of inputs of the RBFNN classifier. Thus the well-known PCA algorithm is applied prior to the RBFNN classification step. In the computer simulation experiments, the RBFNN classifier was trained using pre-classified upper body images and non-person images and then the performance of the proposed classifier for upper body detection is evaluated by using test images and video sequences.

• The KIPS Transactions:PartB
• /
• 제10B권5호
• /
• pp.561-566
• /
• 2003

This paper proposes a new algorithm that generates smooth and realistic transition views from one viewpoint to another view point on the panorama based navigation system. The proposed algorithm is composed with two steps. One is prewarping that aligns the viewing direction in two panorama image, the other is bidirectional disparity morphing (BDM) that generates the intermediate scene from the aligned panorama images. For the prewarping, first of all, we compute the phase correlation between two images in order to get the information such as the displacement, rotation, and scale. Then we align the original images using these information. As soon as finishing the prewarping, we compute the block based disparity vectors (DVs) and smooth them using the two occluding patterns. As we apply these DVs to the BDM, we can get the elaborate intermediate scenes. We make an experiment on the proposed algorithm with real panoramic images and we can get the satisfactory results.

• Journal of Korean Tunnelling and Underground Space Association
• /
• 제16권4호
• /
• pp.387-402
• /
• 2014

In congested urban areas, constructions of tunnel structures have became necessary due to a lack of surface space. The excavation of any tunnel generated the ground disturbances of surrounding ground and displacements is major concern. Therefore, a study of tunnel stability is necessary. In this study, the authors have investigated the stability and failure pattern of tunnel through the model tunnel test. In this study, the close range photogrammetry was used to measure the ground deformation. The measured data was converted to displacement vectors and contours. And then it compared to FE analysis and empirical formula. In addition, this study presented the comparison between steel pipe reinforced model tunnel and unreinforced model tunnel. The ground deformation for both the steel pipe reinforced model tunnel and the unreinforced model tunnel was analysed.

• Journal of the Institute of Electronics and Information Engineers
• /
• 제51권7호
• /
• pp.190-200
• /
• 2014

The video-based system of the broadcasters and the video-related institutions have shifted from analogical to digital in worldwide. This migration process can generate a defect, digital dropout, in the quality of the contents. Moreover, there are limited researches focused on these kind of defects and those related have limitations. For that reason, we are proposing a new method for feature extraction emphasizing in the peculiar block pattern of digital dropout based on discrete cosine transform (DCT). For classification of error block, we utilize support vector machine (SVM) which can manage feature vectors efficiently. Further, the proposed method overcome the limitation of the previous one using continuity of frame by frame. It is using only the information of a single frame and works better even in the presence of fast moving objects, without the necessity of specific model or parameter estimation. Therefore, this approach is capable of detecting digital dropout only with minimal complexity.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• 제39권2호
• /
• pp.116-127
• /
• 2002

A new approach to improve the speaker adaptation algorithm by means of the variable number of observation density functions for CDHMM speech recognizer has been proposed. The proposed method uses the observation density function with more than one mixture in each state to represent speech characteristics in detail. The number of mixtures in each state is determined by the number of frames and the determinant of the variance, respectively. The each MAP Parameter is extracted in every mixture determined by these two methods. In addition, the state segmentation method requiring speaker adaptation can segment the adapting speech more Precisely by using speaker-independent model trained from sufficient database as a priori knowledge. And the state duration distribution is used lot adapting the speech duration information owing to speaker's utterance habit and speed. The recognition rate of the proposed methods are significantly higher than that of the conventional method using one mixture in each state.

• The Journal of the Acoustical Society of Korea
• /
• 제20권1호
• /
• pp.50-55
• /
• 2001

This paper presents a korean isolated words recognition algorithm which used new endpoint detection method, auditory model, 2D-DCT and new distance measure. Advantages of the proposed algorithm are simple hardware construction and fast recognition time than conventional algorithms. For comparison with conventional algorithm, we used DTW method. At result, we got similar recognition rate for speaker dependent korean isolated words and better it for speaker independent korean isolated words. And recognition time of proposed algorithm was 200 times faster than DTW algorithm. Proposed algorithm had a good result in noise environments too.

• Journal of the Korean Data and Information Science Society
• /
• 제23권2호
• /
• pp.285-298
• /
• 2012

This research deals with a relationship between the performance of Korean professional baseball players and their annual salaries. It is based on the sabermetrics, which measures the performance of baseball batters in a refined way. We collect the records of batters of eight professional baseball clubs during the season 2009 and 2010. Then, we calculate every index of the sabermetrics. Principal component analysis is used for examining the relationship between those indexes of sabermetrics and annual salary for the next year. In general, batters who show higher performance get more salary. The result of this research can be useful in order to reach an agreement on salary between a player and his club partner.

• Journal of KIISE:Databases
• /
• 제31권6호
• /
• pp.699-709
• /
• 2004

XML is increasingly important in data exchange and information management. Starting point for retrieving the structure and integrating the documents efficiently is clustering the documents that have similar structure. The reason is that we can retrieve the documents more flexible and faster than the method treating the whole documents that have different structure. Therefore, in this paper, we propose the similar structure-based incremental clustering method useful for retrieving the structure of XML documents and integrating them. As a novel method, we use a clustering algorithm for transactional data that facilitates the large number of data, which is quite different from the existing methods that measure the similarity between documents, using vector. We first extract the representative structures of XML documents using sequential pattern algorithm, and then we perform the similar structure based document clustering, assuming that the document as a transaction, the representative structure of the document as the items of the transaction. In addition, we define the cluster cohesion and inter-cluster similarity, and analyze the efficiency of the Proposed method through comparing with the existing method by experiments.