• Proceedings of the Korean Information Science Society Conference
• /
• 2001.10b
• /
• pp.463-465
• /
• 2001

본 논문에서는 한꺼번에 처리하기 힘든 대용량의 데이터 파일을 전처리 과정을 통해 처리시간을 단축시켜 레이다 영상으로 나타내고 그 안에서 움직이는 물체, 즉 배의 움직임을 추출하고자 한다. 레이더 영상은 일반 직교 좌표계가 아닌 평면상에서의 극좌표계를 사용하기 때문에 좌표계 변환시 생기는 레이다 영상에서의 잡영을 모폴로지 기법과 labeling을 이용하여 제거했다. 레이더 영상은 레이더 자체의 특성으로 인해 잡영이 다수 존재하고 데이터 변화로 인해 영상 전체가 약간의 흔들림이 있기 때문에 물체 추적에 널리 사용되는 차영상 기법을 레이다 영상에 적응시켰을 경우 원하는 물체를 효율적으로 추출해 내기가 어렵다. 따라서 기존의 차영상 추출 알고리즘을 보완하여 이미지 픽셀차가 적정수준 보다 큰 경우만을 선택하여 움직이는 배를 추출하고자 한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2009.11a
• /
• pp.521-522
• /
• 2009

본 논문에서는 이동단말기로부터 수집 가능한 컨텍스트 정보를 기반으로 베이지안 네트워크 추론을 통해 송신자의 정황 정보를 생성하는 시스템을 제안한다. 축적된 데이터로부터 학습되는 베이지안 네트워크의 특성에 따라 설문조사를 통해 사용자의 정황 판단 기호를 수집하고 이를 기반으로 훈련 데이터를 생성하여 베이지안 네트워크를 구성한다. 추론 결과에 대한 사용자 피드백을 주기적인 학습에 사용하고 각 단계에서 정확도를 측정함으로써 개인화된 정황 정보 추론과 사용자의 정황 판단 기호 변화에 신속하게 적응함을 확인한다.

• Phonetics and Speech Sciences
• /
• v.12 no.2
• /
• pp.29-37
• /
• 2020

Recently, the neural network-based deep learning algorithm has dramatically improved performance compared to the classical Gaussian mixture model based hidden Markov model (GMM-HMM) automatic speech recognition (ASR) system. In addition, researches on end-to-end (E2E) speech recognition systems integrating language modeling and decoding processes have been actively conducted to better utilize the advantages of deep learning techniques. In general, E2E ASR systems consist of multiple layers of encoder-decoder structure with attention. Therefore, E2E ASR systems require data with a large amount of speech-text paired data in order to achieve good performance. Obtaining speech-text paired data requires a lot of human labor and time, and is a high barrier to building E2E ASR system. Therefore, there are previous studies that improve the performance of E2E ASR system using relatively small amount of speech-text paired data, but most studies have been conducted by using only speech-only data or text-only data. In this study, we proposed a semi-supervised training method that enables E2E ASR system to perform well in corpus in different domains by using both speech or text only data. The proposed method works effectively by adapting to different domains, showing good performance in the target domain and not degrading much in the source domain.

• 한국HCI학회:학술대회논문집
• /
• 2009.02a
• /
• pp.719-724
• /
• 2009

Fur rendering is difficult in that there are huge numbers of objects and it takes so much time. The previous method considers fur as cylinder, transforms it into 2D ribbon, triangulates and commits rendering. But this method has problem like under sampling and takes rendering time so long. To resolve these shortcuts we proposed new algorithm. We divide fur into thick and thin fur and we applied adaptive rendering methods for each type of fur. Also we can perform an effective rendering according to the proposed rendering framework.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.10a
• /
• pp.612-614
• /
• 2022

This paper propose a clustering algorithm for mobile nodes that possible more efficient clustering using context-aware attribute information in adaptive learning. In typically, the data will be provided to classify interrelationships within cluster properties. If a new properties are treated as contaminated information in comparative clustering, it can be treated as contaminated properties in comparison clustering. In this paper, To solve this problems in this paper, we have new present a context-awareness learning based model that can analyzes the clustering attributed parameters from the node properties using accumulated information properties.

• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.4
• /
• pp.179-188
• /
• 2022

There have been many successful researches on a bearing fault diagnosis based on Deep Learning, but there is still a critical issue of the data distribution difference between training data and test data from their different working conditions causing performance degradation in applying those methods to the machines in the field. As a solution, a data adaptation method has been proposed and showed a good result, but each and every approach is strictly limited to a specific applying scenario or presupposition, which makes it still difficult to be used as a real-world application. Therefore, in this study, we have proposed a method that, using a data transformation with MFCCs and a simple CNN architecture, can perform a robust diagnosis on a target domain data without an additional learning or tuning on the model generated from a source domain data and conducted an experiment and analysis on the proposed method with the CWRU bearing dataset, which is one of the representative datasests for bearing fault diagnosis. The experimental results showed that our method achieved an equal performance to those of transfer learning based methods and a better performance by at least 15% compared to that of an input transformation based baseline method.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.05a
• /
• pp.646-648
• /
• 2021

Sensors with wireless communication functions are essential for acquiring and transmitting spatio-temporal data that is not easily accessible in sensor network environments. However, these sensors lack adaptability to large amounts of sensing data processing or dynamic environments, resulting in over-consumption of power and network overhead. This paper proposes a mobile agent that can acquire, transmit, and process only the necessary data by applying thresholds, and presents methods for autonomous migration and communication processing of mobile agents.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.40 no.3
• /
• pp.506-515
• /
• 2015

In this paper, we propose a group playback synchronization method using adaptive synchronization delay time by the bit rate of media to synchronize a play position of streaming media between mobile smart devices. This method consists of streaming server-side and client-side synchronization algorithms based on synchronization delay time which includes connection time, control packet transmission time, streaming data buffering time, and synchronization processing time. We implement the Android media player application with synchronization support using the proposed algorithms and present the result of performance evaluation.

• The KIPS Transactions:PartA
• /
• v.15A no.3
• /
• pp.151-154
• /
• 2008

A middleware in grid computing environment is required to support seamless on-demand services over diverse resource situations in order to meet various user requirements [1]. Since grid computing applications need situation-aware middleware services in this environment. In this paper, we propose a semantic middleware architecture to support dynamic software component reconfiguration based fault and service ontology to provide fault-tolerance in a grid computing environment. Our middleware includes autonomic management to detect faults, analyze causes of them, and plan semantically meaningful strategies to recover from the failure using pre-defined fault and service ontology trees. We implemented a referenced prototype, Web-service based Application Execution Environment(Wapee), as a proof-of-concept, and showed the efficiency in runtime recovery.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2008.11a
• /
• pp.68-71
• /
• 2008

웹 상에 정보가 폭발적으로 증가함에 따라 각 사용자에게 맞는 정보를 선별하여 제공하는 개인화 서비스는 매우 중요한 이슈가 되었다. 기존 추천시스템들은 컨텐츠 기반 필터링과 협업 필터링 기법을 기반으로 한다. 그러나 이러한 방법들은 충분히 수집된 사용자 정보를 필요로 하기 때문에, 적절한 추천이 이루어지기 까지 다소 시간이 소요되는 문제를 가지고 있다. 또한 사용자의 성향이 지나치게 편중되는 경우, 사용자의 취향변화를 반영하여 새로운 상품을 추천하는 것은 어렵다. 실제로 사용자들은 웹 사이트의 방문 목적에 따라 개인화된 상품추천을 원하기도 하고, 많은 사용자들에게 인기 있는 상품을 원하기도 한다. 본 논문에서는 사용자의 행동분석을 기반으로, 협업 필터링을 기반으로 하는 개인화된 추천과 다수의 사용자들에게 공통적으로 인기 있는 상품의 추천 비율을 동적으로 조합하여 최종 추천 상품들을 선별하는 새로운 적응형 추천 시스템을 제안한다. 본 논문에서는 MovieLens의 데이터 셋을 이용하여 기존 추천기법들과 추천결과에 대한 정확도를 비교 실험하였으며, 보다 높은 정확도를 보이는 실험결과를 통해 제안시스템의 유효성을 확인하였다.