• 한국음향학회지
• /
• 제27권7호
• /
• pp.379-385
• /
• 2008

본 논문에서는 근단 (Hear-End)및 원단 (Far-End) 잡음 환경에서 효과적인 음성 강화 기법을 제시한다. 일반적으로 배경 잡음이 존재하는 근단 환경에서 수신하는 원단 화자 음성의 명료도가 매우 감소하므로, 이를 극복하기 위한 원단 화자 음성 강화 기법이 필요하다. 구체적으로, 추정된 근단 화자의 배경 잡음 전력을 기반으로 원단 화자의 음성 전력을 강화시키는데, 특별히 근단 환경에서도 잡음이 존재하는 일반적인 경우를 고려하여, 잡음에 오염된 원단 음성 신호중 잡음을 제외한 실제 음성 신호만 강화하는 개선된 알고리즘을 제안한다 제안된 음성 강과 기법의 성능은 다양한 잡음 환경 하에서 ITU-T P.800의 주관적 음질 측정 방법인 CCR (Comparison Category Rating) 테스트에 의해 평가되었으며, 기존의 음성 강화기법과 비교해서 우수한 성능을 보여주었다.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2003년도 ICCAS
• /
• pp.2445-2450
• /
• 2003

Recently, biological technology industry shows great development. Instruments and systems related biological technology have been developed actively. In this paper, we developed a new micro end-effector for biological cell manipulation. The existing micro end-effector for biological cell manipulation has not any force sensing mechanism. Usually, excessive contact force occurring when the end-effector and a cell collide might make a damage on the cell. However, unfortunately, user can not notice the condition in case of using the existing end-effector. In order to overcome we proposed the improved micro end-effector having a force sensing mechanism. This paper presents the design concepts of the new micro end-effector. We carried out calibration of the force sensor and tested the performance of the proposed micro end-effector. Through a series of experiments the new micro end-effector shows the possibility of application for precision biological cell manipulation such as DNA operation

• 말소리와 음성과학
• /
• 제10권1호
• /
• pp.39-48
• /
• 2018

A typical statistical parametric speech synthesis (text-to-speech, TTS) system consists of separate modules, such as a text analysis module, an acoustic modeling module, and a speech synthesis module. This causes two problems: 1) expert knowledge of each module is required, and 2) errors generated in each module accumulate passing through each module. An end-to-end TTS system could avoid such problems by synthesizing voice signals directly from an input string. In this study, we implemented an end-to-end Korean TTS system using Google's Tacotron, which is an end-to-end TTS system based on a sequence-to-sequence model with attention mechanism. We used 4392 utterances spoken by a Korean female speaker, an amount that corresponds to 37% of the dataset Google used for training Tacotron. Our system obtained mean opinion score (MOS) 2.98 and degradation mean opinion score (DMOS) 3.25. We will discuss the factors which affected training of the system. Experiments demonstrate that the post-processing network needs to be designed considering output language and input characters and that according to the amount of training data, the maximum value of n for n-grams modeled by the encoder should be small enough.

• 한국정보과학회 언어공학연구회:학술대회논문집(한글 및 한국어 정보처리)
• /
• 한국정보과학회언어공학연구회 2017년도 제29회 한글 및 한국어 정보처리 학술대회
• /
• pp.111-115
• /
• 2017

목적 지향적 대화 시스템(Goal-oriented dialogue system) 은 텍스트나 음성을 통해 특정한 목적을 수행 할 수 있는 시스템이다. 최근 RNN(recurrent neural networks)을 기반으로 대화 데이터를 end-to-end learning 방식으로 학습하여 대화 시스템을 구축하는데에 활용한 연구가 있다. End-to-end 방식의 학습은 도메인에 대한 지식 없이 학습 데이터 자체만으로 대화 시스템 구축을 위한 학습이 가능하다는 장점이 있지만 도메인 지식을 학습하기 위해서는 많은 양의 데이터가 필요하다는 단점이 존재한다. 이에 본 논문에서는 도메인 특정 지식을 결합하여 end-to-end learning 방식의 학습이 가능한 Hybrid Code Network 구조를 기반으로 한국어로 구성된 식당 예약에 관련한 대화 데이터셋을 이용하여 식당 예약을 목적으로하는 대화 시스템을 구축하는 방법을 제안한다. 실험 결과 본 시스템은 응답 별 정확도 95%와 대화 별 정확도 63%의 성능을 나타냈다.

• 한국가금학회지
• /
• 제24권1호
• /
• pp.29-37
• /
• 1997

An experiment was conducted for 20 days( from July 21 to August 10, 1996) to investigate the differences in thickness of eggshell and eggshell membrane in Korean native chicken. All the eggs were purchased from a farm located in Chinju area. All eggs were categorized into Treatment I (34.89~51.39 g), H (53.32~60.70 g), and III (34.89~60.70 g) according to their weights. Eleven items were measured in all three treatments, i.e., egg weight, egg length, egg width, breaking strength, eggshell thickness of large end(STLE), eggshell membrane thickness of large end(SMTL), eggshell thickness of small end (STSE), eggshell membrane thickness of small end(SMTS), eggshell thickness of middle part(STMP), and eggshell membrane thickness of middle part(SMTM). Correlation coefficients between egg weight and eggshell weight, among eggshell thickness of large end, small end and middle end and eggshell weight were significant (P<0.01). Correlation coefficients between egg weight and large eggshell membrane were nega-tively significant(P<0.01). Breaking strengths were higher in Treatment I than in Treatment H. Correlation coefficients among thickness of eggshell and eggshell membrane at each point at 3 parts of large end, small end and middle end were not significant.

• 한국정보시스템학회지:정보시스템연구
• /
• 제10권1호
• /
• pp.73-95
• /
• 2001

A significant body of information systems research is concerned with user participation as a means to improve user satisfaction within systems development. In spite of the many studies in this area, findings are not consistent or cumulative. One of the reasons for this is not to be considered the type of systems. Most research are focusing on traditional system development under the mainframe environment instead of under end-user based environment like web-based information system or ERP. End-user based environment requires user's strong initiatives in the system development life cycle. It means the user\`s simple participation is not enough and user-led-development is more appropriate for the system success. This paper can call this new concept is end-user joint work which means user has to participate in the system development as a co-developer. This paper investigates how the end-user joint work contributes end-user satisfaction and two variables in particular-job complexity and technical competence-moderate the outcomes of end-user joint work Analysis of 112 independent systems development projects indicated that end-user joint work has a direct relationship with end-user satisfaction. In addition, the two variables were found to moderate key roles on this relationship. That is, the strength of this relationship depended on the level of these factors.

• International Journal of Precision Engineering and Manufacturing
• /
• 제4권4호
• /
• pp.19-24
• /
• 2003

In end milling process, which is characterized by the use of a rotating tool, the undeformed chip thickness varies periodically with phase change of the tool. Although many efforts have concentrated on the study of end milling process, the analysis of shear and chip-tool friction behaviors has not been reported. Recently, a model has been proposed to simulate the shear and friction characteristics of an up-end milling process in terms of the equivalent oblique cutting. In the current study, the varying undeformed chip thickness and the cutting forces in a down-end milling process are replaced with the equivalent ones of oblique cutting. Then it is possible to simulate the shear and the chip-tool friction characteristics of a down-end milling process. The proposed model has been verified through two sets of cutting tests i.e., down-end milling and the equivalent oblique cutting tests. The experimental results show that the proposed model is suitable to analyze the shear and chip-tool frictional characteristics of down-end milling process. The specific cutting energy decreases with increase in equivalent undeformed chip thickness in a down-end milling process.

• 한국소성가공학회:학술대회논문집
• /
• 한국소성가공학회 2005년도 추계학술대회 논문집
• /
• pp.344-349
• /
• 2005

This paper is studied on the effect of TiAlN coated ball end mill on cutting characteristic of high hardness steels in dry cutting condition without coolant. KP4 steels[HRC32] and STD11[HRC60] heat treated steels wert: used as the workpiece and WC-Co ball end mill and single and multi layer TiAlN coated ball end mill were utilized. Results showed that TiAlN coated ball end mill were increased the cutting length than WC-Co ball end mill in the cutting speed$[245\~320m/min]$ about $2\~5$ times for KP4 steels and about $2.7\~4.3$ times for STD11 heat treated steels. The multi layer TiAlN coated ball end mill is good about $1.2\~1.7$ times for KP4 steels and STD11 steels than single layer coated.

• 한국의류학회지
• /
• 제30권1호
• /
• pp.12-19
• /
• 2006

This study was administered to identify features of masstiege high-end products by comparing to original old high-end products. For this purpose, the differences in shopping orientation between each consumer groups by product types(high-end products, original old high-end products) preferred were examined. 300 female consumers were surveyed and 279 data were used for analysis. The results are as follows. First, three elements(efficiency, enjoyment, convenience) of shopping orientation were constructed by factor analysis and efficiency element of all elements had highest explanation power. The differences in shopping orientation between masstiege high-end products and original old high-end products were identified by t-test. The consumers preferring original old high-end products regarded convenience element as an important factor and the consumers preferring masstiege high-end products considered efficiency element more. The correlation analysis between shopping orientation factors and demographic characteristics were administered. The consumers who were of low age, low education, low income and unmarried pursued efficiency more, the consumers who were unmarried, of high education and high income showed to pursue enjoyment more, and the consumers who were of high age, high education, high income and married considered convenience element more. Finally marketing strategies for masstiege brands were suggested.

• 한국멀티미디어학회논문지
• /
• 제24권5호
• /
• pp.676-683
• /
• 2021

Wi-Fi Round-trip timing (RTT) based location estimation technology estimates the distance between the user and the AP based on the transmission and reception time of the signal. This is because reception instability and signal distortion are greater than that of a Received Signal Strength Indicator (RSSI) based fingerprint in an indoor NLOS environment, resulting in a large position error due to multipath fading. To solve this problem, in this paper, we propose an end-to-end based WiFi Trilateration Net (WTN) that combines neural network-based RTT correction and trilateral positioning network, respectively. The proposed WTN is composed of an RNN-based correction network to improve the RTT distance accuracy and a neural network-based trilateral positioning network for real-time positioning implemented in an end-to-end structure. The proposed network improves learning efficiency by changing the trilateral positioning algorithm, which cannot be learned through differentiation due to mathematical operations, to a neural network. In addition, in order to increase the stability of the TOA based RTT, a correction network is applied in the scanning step to collect reliable distance estimation values from each RTT AP.