• Proceedings of the KSPS conference
• /
• 2006.05a
• /
• pp.139-142
• /
• 2006

Speaker verification systems can be implemented using speaker adaptation methods if the amount of speech available for each target speaker is too small to train the speaker model. This paper shows experimental results using well-known adaptation methods, namely Maximum A Posteriori (MAP) and Maximum Likelihood Linear Regression (MLLR). Experimental results using Korean speech show that MLLR is more effective than MAP for short enrollment utterances.

• The Journal of Korean Institute for Practical Engineering Education
• /
• v.2 no.1
• /
• pp.1-7
• /
• 2010

Recently, interest in u-learning to pursue effective learning by using ubiquitous environment in teaching and learning activities. In u-learning environment, learners should be able to push necessary information at the right time and the right place. Also calm technology oriented to, and this means that it can recognize learners' terminal information and to provide adaptive interface. In u-learning environment, main learning terminals would be mobile terminals which support mobility. However, learning in the existing PC environment should not be excluded. Thus, by providing adaptive interface according to various learners' terminal in LMS for u-learning, learners are able to learn through consistent and natural learning interface with any computer or any network at any place and at any time. The purpose of this study is to propose the interface adaptation based on terminal information focusing on the layout transformation process in the development environment.

• Proceedings of the KSPS conference
• /
• 2005.11a
• /
• pp.159-162
• /
• 2005

Identification is the process automatically identify who is speaking on the basis of information obtained from speech waves. In training phase, each speaker models are trained using each speaker's speech data. GMMs (Gaussian Mixture Models), which have been successfully applied to speaker modeling in text-independent speaker identification, are not efficient in insufficient training data environment. This paper proposes speaker modeling method using MLLR (Maximum Likelihood Linear Regression) method which is used for speaker adaptation in speech recognition. We make SD-like model using MLLR adaptation method instead of speaker dependent model (SD). Proposed system outperforms the GMMs in small training data environment.

• Proceedings of the KIEE Conference
• /
• 2004.05a
• /
• pp.68-70
• /
• 2004

In this paper we develop an adaptive interface between video compression and transission protocols to handle QoS fluctuations that are common to mobile communication systems. We consider various generic design alternatives for QoS adaptation and identify 'QoS negotiation' as the most promissing. This method gives the best possibilities to obtain system-wide efficiency. To handle the indued system complexity we apply a design philosophy (called ARC) that separates implementation dependencies by introducing QoS interfaces between system modules. In the ARC phlosophy the implementation details are hidden in the subsystems. To assure efficient adaptation, the QoS must be negotiated between modules. We select the QoS parameters that are both necessary and sufficient for efficient negotiation between the video encoder and protocol modules. We describe the relation between the QoS parameters at the interface and the internal parameters of common video coding methods and protocol elements. Furthermore, we describe a negotiation procedure that allows a system-wide optimum to emerge.

• Restorative Dentistry and Endodontics
• /
• v.19 no.2
• /
• pp.641-654
• /
• 1994

This study was executed to evaluate adaptability of dentin bonding agents to dentinal wall with measuring contraction gap on interface between cavity wall and composite resin by SEM study. 6 kinds of dentin bonding agents were investigated for this study : Gluma, Super Bond C & B, All bond 2, Scotchbond multipurpose, Scotchbond 2 and Clearfil photo bond. 30 of fresh extracted teeth were randomly selected and divided into 6 groups with each 5. The round shaped cavities with 3mm dia. and 1.5mm depth were prepared on cementoenamel junction of buccal surface of teeth. Dentin bonding agents were applied to cavity wall and then the composite resin was filled in the cavity. Specimens were sectioned longitudinally on buccal surface. Sectioned aspects of specimens were impressioned with rubber base materials and finally precise replica were made of epoxy resin poured in negative impression. Contraction gaps were examined on interface between cavity wall and composite resin under condition of 200 and 2000 magnification of SEM. The results were as follows. 1. There were no gap on interface between enamel and composite resin in all specimens, but gaps were mainly exhibited on apical side of lateral wall of dentin of cavity. 2. In Gluma, 2 cases of 5 specimens exhibited excellent adaptation to the cavity wall, indicating no gaps on interface between cavity wall and composite resin. The other specimens showed gaps with range of $0{\sim}15{\mu}m$ width. 3. In Super Bond C & B, gaps with range of $0{\sim}10{\mu}m$ width were mainly exhibited on apical side of lateral wall of dentin of cavity. 4. In All bond 2, all specimens showed the most exellent adaptation to cavity wall when compared to the other materials, indicating no gap interface between cavity wall and composite resin. 5. In Scotchbond multipurpose, gaps with range of $0{\sim}10{\mu}m$ width were locally located on cavity wall. 6. In Scotchbond 2, all specimens showed great amount of gap with range of $5{\sim}25{\mu}m$, indicating the worst adaptation to cavity wall compared to the other materials. 7. In Clearfil photo bond, 2 case of 5 specimens exhibited exellent adaptation to cavity wall, indicating no gap. the other specimens showed gap with range of $0{\sim}15{\mu}m$ width on inferface between cavity wall and composite resin.

• Restorative Dentistry and Endodontics
• /
• v.13 no.1
• /
• pp.139-150
• /
• 1988

The purpose of this study was to evaluate the marginal adaptation of Class V resin restorations by the filling techniques. The cavities were filled with Heliosit$^{(R)}$ light curing composite resin system in accordance with the following three filling techniques - bulk-pack placement technique, apical to occlusal layering placement technique, and buccolingual layering placement technique. And the interface between the restoration and the tooth structure was observed with scanning electron microscope. The following results were obtained. 1. Marginal adaptation of enamel showed excellent irrespective of the filling techniques. 2. Marginal leakage of dentin was observed irrespective of the filling techniques. 3. No difference of marginal adaptation according to the filling techniques was observed in light curing composite resin.

• Journal of Broadcast Engineering
• /
• v.10 no.2
• /
• pp.139-148
• /
• 2005

This paper presents recent works of the development of MPEG-21 DIA utility software as one of MPEG-21 activities. The main purpose of this work is to make integrated utility software for DIA C/C++ modules in a Java framework. To do this, JNI plays a role of the interface of a main module with C/C++ DIA adaptation modules. Furthermore, DIA and DID parsers are integrated to the utility software. Our software is designed such that all DIA modules are easily integrated. A variety of adaptation modules have been adopted in MPEG-21 standard. Among them, stereoscopic video conversion that is one of DIA adaptation modules is chosen and we show that our utility software is correctly implemented based on experiments.

• Restorative Dentistry and Endodontics
• /
• v.18 no.2
• /
• pp.413-422
• /
• 1993

The purpose of this study was to evaluate the adaptation of light cured glass ionomer cement and composite resin using all- etch technique to tooth structure. In this study, class V cavities were prepared on the buccal surfaces of 10 extracted human premolar teeth with cementum margin and teeth were randomly assigned 2 groups of 5 teeth each. The cavities of glass ionomer cement group were filled with the light cured glass ionomer cement(Fuji II LC) and the cavities of composite resin group were filled with the light cured composite resion(P - 50) using all- etch technique with All- Bond 2. The restored teeth were stored in 100 % relative humidity at $37^{\circ}C$ for 48 hours. And then, the roots of the teeth were removed with the tapered fissure bur and the remaining crowns were sectioned occlusogingivally through the center of restorations. Adaptation at tooth - restoration interface were assessed occlusally, gingivally, and axially by scanning electron microscope. The results were as follows : 1. The adaptation to enamel walls of composite resin restorations using All - Bond 2 showed better than glass ionomer restorations. 2. The adaptation to gingival and axial walls of glass ionomer restorations showed better than composite resin restorations using All - Bond 2. 3. In both groups, occlusal margins of restorations showed better adaptation than gingival margins of restorations.

• Journal of Institute of Control, Robotics and Systems
• /
• v.10 no.5
• /
• pp.464-474
• /
• 2004

This paper proposes a middleware structure for the module-based personal robot, which can run on heterogeneous network interfaces and provides users easy interface-method regardless of underlying heterogeneous interfaces and convenient exchange of modules. The proposed middleware is divided into three layers of a streaming layer (SL), a network adaptation layer (NAL) and a network interface layer (NIL). The streaming layer manages application transactions using middleware services and provides user a uniform interfaces to the proposed middleware. The network adaptation layer manages a message-routing and provides naming service and it is a core of the proposed middleware. And the network interfaces layer manages dependent parts of heterogeneous network interfaces such as IEEE1394, USB, Ethernet, and CAN (Control Area Network). This paper implements the proposed middleware structure, where 3 types of interfaces of IEEE 1394, USB and Ethernet are used, and measures response times among those interfaces.

• Journal of the Korean Society for Aviation and Aeronautics
• /
• v.7 no.1
• /
• pp.7-18
• /
• 1999

The FAA, along with the ICAO and other members of the civil aviation community, has recognized that a GNSS will provide the primary stand-alone navigation system in the 21st Century. FAA has initiated plans to transition from its present ground-based navigation and landing system to satellite-based using signals generated by the GPS. In spite of some risks, GPS users are increasing rapidly. About 52 aircraft equipped with various GPS in their system and wide spread of GPS may be expected in Korea. However, the regulations concerning with CPS implementation were not established by the government. Another problem is GPS receiver's interface. The user interface, operating method and capability vary with GPS class and model. As a direct operator for the system, pilots have to ensure these limitations and rules for efficient adaptation and safety. The issues identified by the study are highly interrelated, and are evidence of aviation system problem. To treat one issue in isolation may improve certain aspects of the aviation system, but will ultimately fail to fundamentally increase the safety and efficiency for the system.