• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.45 no.2
• /
• pp.37-48
• /
• 2008

Generally, a motion vector and a disparity vector represent the motion information of an object in a single-view of camera and the displacement of the same scene between two cameras that located spatially different from each other, respectively. Conventional H.264/AVC does not use the disparity vector in the motion vector prediction because H.264/AVC has been developed for the single-view video. But, multi-view video coding that uses the inter-view prediction structure based on H.264/AVC can make use of the disparity vector instead of the motion vector when the current frame refers to the frame of different view. Therefore, in this paper, we propose an improved motion/disparity vector prediction method that consists of global disparity vector replacement and extended neighboring block prediction. From the experimental results of the proposed method compared with the conventional motion vector prediction of H.264/AVC, we achieved average 1.07% and 1.32% of BD (Bjontegaard delta)-bitrate saving for ${\pm}32$ and ${\pm}64$ of global vector search range, respectively, when the search range of the motion vector prediction is set to ${\pm}16$.

• The Journal of the Acoustical Society of Korea
• /
• v.24 no.8
• /
• pp.498-504
• /
• 2005

We introduce a Korean speech recognition platform (ECHOS) developed for education and research Purposes. ECHOS lowers the entry barrier to speech recognition research and can be used as a reference engine by providing elementary speech recognition modules. It has an easy simple object-oriented architecture, implemented in the C++ language with the standard template library. The input of the ECHOS is digital speech data sampled at 8 or 16 kHz. Its output is the 1-best recognition result. N-best recognition results, and a word graph. The recognition engine is composed of MFCC/PLP feature extraction, HMM-based acoustic modeling, n-gram language modeling, finite state network (FSN)- and lexical tree-based search algorithms. It can handle various tasks from isolated word recognition to large vocabulary continuous speech recognition. We compare the performance of ECHOS and hidden Markov model toolkit (HTK) for validation. In an FSN-based task. ECHOS shows similar word accuracy while the recognition time is doubled because of object-oriented implementation. For a 8000-word continuous speech recognition task, using the lexical tree search algorithm different from the algorithm used in HTK, it increases the word error rate by $40\%$ relatively but reduces the recognition time to half.

• Journal of KIISE:Computing Practices and Letters
• /
• v.12 no.1
• /
• pp.66-77
• /
• 2006

In this paper, we researched the searching and tracking locations of a home resident using ON/OFF switches and sensors and designed a remote monitoring system. As an implementation environment, this system is developed on the base of the distributed object group framework we have developed from previous works. In order to trace the moving locations of a home resident, we firstly showed a home structure which attaches ON/OFF switches and sensors to home appliances and indoor facilities being fixed in home. Whenever a home resident opens/closes these objects, the signals operated from ON/OFF switches and sensors are sent to a home server system. In this time, the real locations of ON/OFF switches and sensors that the signals are being occurred must be the current location that he/she stays. A home server system provides the functionalities that map the real location of a resident in home to virtual location designed on remote desk-tops or terminals like PDAs, and that construct a healthcare database consisted of moving patterns, moving ranges, momentum for analyzing the given searching locations and times Finally, this system provides these information for remotely monitoring services.

• The Transactions of the Korea Information Processing Society
• /
• v.3 no.5
• /
• pp.1262-1272
• /
• 1996

Pure functional languages have the referential transparency feature sothat all objects cannot be updated destuctively. Once an aggregated data structure such as array is updated, both the original and newly updated array must be preserved to maintain reverential transparency. Thus, it is required to develop an efficient mechanism with which can reduce the maintenance cost while maintaining referential transparency for whole data. This study is to suggest a functional array to solve the problem, and them test it. For that, the proposed mechanism was implemented on a combinator graph reduction machine. The result shows that proposed mechanism reduces traversal time for array operations. Also, updating all versions and accessing the recent version are achieved in constant time without reconstruction of updated data in execution time.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2007.10a
• /
• pp.398-403
• /
• 2007

초음파 영상은 미세한 명암도 차이 등에 의해 분석 과정에서 근육 영역의 위치와 크기를 판단하는데 어려움이 발생하고 이로 인해 근육 영역을 파악하는데 주관성이 개입된다. 본 논문에서는 근육영역을 객관적으로 분석하기 위해 ART2 신경망을 적용하여 양자화를 수행한 후, 국부적 영역에서 근육 영역을 추출한다. 초음파 영상에서 히스토그램 평활화와 엔드인 탐색 알고리즘을 적용하여 명암도의 분포와 밝기 값을 보정 한 후, ART2 신경망을 이용하여 유사한 영역을 클러스터링 한다. 그리고 클러스터링 된 각 영역의 크기, 위치 및 명암도 정보를 분석하여 피하지방, 근육 막, 기타 배경 영역으로 분류한다. 최종적인 근육 영역을 찾기 위해 근육 막 내부 객체들 간의 거리, 각도를 이용하여 근육 막 영역에 둘러싸인 근육 영역을 추출한다. 실제 초음파 영상을 대상으로 실험한 결과, 일반적인 클러스터링 기법을 적용한 방법 보다 ART2 기반 양자화와 제안된 영역 확장 기법으로 근육영역을 추출하고 분석하는 것이 효율적임을 확인하였다.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.37 no.6A
• /
• pp.497-502
• /
• 2012

In this paper, we present an immersive videoconferencing system that enables gaze correction between users in the internet protocol TV (IPTV) environment. The proposed system synthesizes the gaze corrected images using the depth estimation and the virtual view synthesis algorithms as one of the most important techniques of 3D video system. The conventional processes, however, causes several problems, especially temporal inconsistency of a depth video. This problem leads to flickering artifacts discomforting viewers. Therefore, in order to reduce the temporal inconsistency problem, we exploit the joint bilateral filter which is extended to the temporal domain. In addition, we apply an outlier reduction operation in the temporal domain. From experimental results, we have verified that the proposed system is sufficient to generate the natural gaze-corrected image and realize immersive videoconferencing.

• Journal of the Korea Society of Computer and Information
• /
• v.19 no.6
• /
• pp.11-17
• /
• 2014

While safe and convenient, ultrasound imaging analysis is often criticized by its subjective decision making nature by field experts in analyzing musculoskeletal system. In this paper, we propose a new automatic method to extract muscle area using ART2 neural network based quantization. A series of image processing algorithms such as histogram smoothing and End-in search stretching are applied in pre-processing phase to remove noises effectively. Muscle areas are extracted by considering various morphological features and corresponding analysis. In experiment, our ART2 based Quantization is verified as more effective than other general quantization methods.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2006.11a
• /
• pp.23-26
• /
• 2006

최근 대부분의 이러닝(E-Learning)을 교육하는 사이트는 학습 콘텐츠를 검색하는 방법이 단순한 리스트의 나열과 택스트 매칭(Text matching)방법을 사용하는 단점이 있다. 이를 보완하기 위해 좀 더 컴퓨터가 정보 데이터의 의미를 분석하여 검색이 가능하도록 개념 네트워크인 시맨틱웹(Semantic Web)이 등장하였다. 본 논문에서는 이러한 시맨틱웹의 온톨로지(Ontology) 언어 중에 토픽맵(Topic Maps)을 사용하여 많은 양의 학습 정보 데이터를 쉽고도 정확하게 연결 지어 학습 콘텐츠에 대한 정보를 표현하고, 구조화할 수 있는 방법을 모색해 보고자 한다. 학습자의 관심분야 정보, 학습객체의 학습 권장자의 정보와 함께 학습 경험과 검색 빈도수를 분석한 협력 필터링과 학습 에이전트의 개인화 기법을 동시에 사용하여 선호도를 분석한다. 이 선호도를 가지고 학습자의 메타데이터를 생성하고, 로그 데이터를 따로 데이터베이스에 저장한다. 이러한 학습자의 정보와 학습 콘텐츠간의 정보를 상호 연결하여, 그 토픽맵을 사용하여 연관관계를 정의해 줌으로써 학업성취도를 높이고, 학습자 개개인의 성향에 가장 알맞은 학습 콘텐츠를 탐색해가는 네비게이터(Navigator)를 설계하였다.

• Journal of KIISE:Information Networking
• /
• v.35 no.5
• /
• pp.367-373
• /
• 2008

Recently, the NAT middlebox widely deployed in the home network environment prohibits DM operations from reaching user devices behind NAT. In this article, we focus on NAT issues to manage home network devices. Particularly, we discuss standardization efforts, and present our proposal to deploy DM services for VoIP and IPTV devices under NAT. By slightly changing behaviors of Simple Network Management Protocol (SNMP) Manager and Agent, and defining additional Management Objects (MOs) to gather NAT binding information, we could solve the NAT traversal problem under symmetric NAT. Moreover, we propose an enhanced method to search the UDP hole binding time of the NAT box. We applied our method to randomly selected 22 VoIP devices out of 194 NATed hosts in the real broadband network and have achieved 99% of the success ratio for exchanging SNMP request messages and 26% of enhancement for searching the UDP hole binding time.

• Korean Journal of Remote Sensing
• /
• v.19 no.4
• /
• pp.319-328
• /
• 2003

The automatic analysis of high-resolution satellite image is important in cartography, surveillance, exploiting resources etc. However, the automatic analysis of high resolution satellite image in the urban area has lots of difficulty including a shadow, the difference of illumination with time, the complexity of image so that the present techniques are seemed to be impossible to resolve. This paper proposes a new way of change detection of building objects in urban area, in which the objects in digital vector map are designated and superimposed on the the high-resolution satellite image. The proposed way makes the buildings on the vector map parameterize, and searches them in the preprocessed high-resolution satellite image by using generalized Hough transform. The designated building objects are overlaid on the satellite image and the result can help to search the changes in building objects rapidly.