• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2003년도 춘계학술발표논문집 (상)
• /
• pp.567-570
• /
• 2003

본 논문에서는 연속영상 속에 있는 얼굴영역을 칼만 필터를 이용하여 추적하는 방법을 제안한다. 제안된 방법은 영역-기반접근(region-based approach)방법인 워터쉐드 알고리즘을 이용하여 초기 영역 분할 작업을 한 후 얼굴칼라 모델과의 매칭작업을 통해서 얼굴영역을 찾아내는 얼굴검출 단계와 추출된 얼굴영역의 칼라정보를 칼만 필터의 입력으로 하여 얼굴을 추적하는 단계로 구성되어 있다. 실험결과를 통하여 제안된 방법이 배경이 복잡한 영상에 대해서도 효율적으로 얼굴을 추적할 수 있음을 보인다.

• Earthquakes and Structures
• /
• 제12권5호
• /
• pp.569-581
• /
• 2017

The Monte Carlo Simulation (MCS) based seismic fragility analysis (SFA) approach allows defining more realistic relationship between failure probability and seismic intensity. However, the approach requires simulating large number of nonlinear dynamic analyses of structure for reliable estimate of fragility. It makes the approach computationally challenging. The response surface method (RSM) based metamodeling approach which replaces computationally involve complex mechanical model of a structure is found to be a viable alternative in this regard. An adaptive moving least squares method (MLSM) based RSM in the MCS framework is explored in the present study for efficient SFA of existing structures. In doing so, the repetition of seismic intensity for complete generation of fragility curve is avoided by including this as one of the predictors in the response estimate model. The proposed procedure is elucidated by considering a non-linear SDOF system and an existing reinforced concrete frame considered to be located in the Guwahati City of the Northeast region of India. The fragility results are obtained by the usual least squares based and the proposed MLSM based RSM and compared with that of obtained by the direct MCS technique to study the effectiveness of the proposed approach.

• ETRI Journal
• /
• 제29권5호
• /
• pp.700-702
• /
• 2007

We present a relevance feedback approach based on multi-class support vector machine (SVM) learning and cluster-merging which can significantly improve the retrieval performance in region-based image retrieval. Semantically relevant images may exhibit various visual characteristics and may be scattered in several classes in the feature space due to the semantic gap between low-level features and high-level semantics in the user's mind. To find the semantic classes through relevance feedback, the proposed method reduces the burden of completely re-clustering the classes at iterations and classifies multiple classes. Experimental results show that the proposed method is more effective and efficient than the two-class SVM and multi-class relevance feedback methods.

• The Journal of Asian Finance, Economics and Business
• /
• 제4권4호
• /
• pp.5-15
• /
• 2017

The Northern States of Malaysia comprises of four states (Penang, Kedah, Perlis and Perak) still record high poverty incidence eventhough Malaysia has experienced a remarkable reduction of poverty over the past century. Economic activities in Perlis and Kedah that are predominantly agriculture in the rural area contribute to this disparity. To add, rural households are also subject to risks and uncertainties that make them more vulnerable to poverty. This study examines the impact of risks and assets on households' vulnerability to poverty. A survey of 400 respondents was conducted in December 2015 in the northern region of Malaysia. From these 400 questionnaires, only 298 were considered valid and used in the analysis. Using a logistic probability function, the results indicated that risks are not a significant threat to households. Gender and strata are crucial elements that significantly determine households' vulnerability. While human capital and financial capital significantly reduce households' vulnerability to poverty, physical and natural capitals were not statistically significant. The study suggests that the government and practitioners design strategies and policies with an assets-based approach. The asset-based approach is more appropriate for linking the causes of poverty to vulnerability.

• 정보처리학회논문지B
• /
• 제13B권4호
• /
• pp.383-392
• /
• 2006

본 논문에서는 영역기반 영상검색의 성능 향상을 위한 피드백 알고리즘으로 다중 클래스를 갖는 확률적 신경망(Probabilistic Neural Networks)을 이용한 방법론을 제안하고 이를 영역기반 영상 검색 시스템인 FRIP(Finding Regions In the Pictures) 시스템에 적용하였다. 본 논문에서 제안하는 피드백 알고리즘은 특정 벡터가 독립적이라는 가정을 할 필요가 없으며 보다 상세한 분류를 위해 추가적인 클래스들을 추가할 수 있도록 허용하고 있다. 또한 단지 4개 층(layer)만을 가지고 있음으로 학습을 위한 계산시간이 적게 든다는 장점이 있다. 추가적으로 다음단계에서의 성능 향상을 위해 분류 단계에서 사용자의 이전 피드백 행동을 모두 히스토리(history)로 모두 기억시켜 놓고 다음 단계를 위한 가중치 학습을 위해 사용하도록 한다. 히스토리를 사용함으로써 제안하는 알고리즘은 사용자의 주관적 의도를 보다 정확하게 파악 할 수 있을 뿐만 아니라 학습을 위해 이전 단계만을 사용 했을 때 발생할 수 있는 성능 감소를 막을 수 있다. 본 논문에서는 Corel-photo CD에서 3000장의 자연 영상을 무작위로 추출하여 기존의 방법론들과 제안하는 방법론의 성능을 측정하여 본 논문에서 제안하는 방법론이 성능이 우수함을 증명하였다.

• 농촌계획
• /
• 제11권4호
• /
• pp.33-45
• /
• 2005

This study investigates the core elements of the formation and development of cluster using a multi-disciplinary approach and suggests a promoting strategy by development stage of cluster. As a sub-category of regional innovation system, the cluster has been considered as one of the most noticeable methodological argument to make the regional innovation system come true. In the meantime, this study examines the core elements of cluster shown in the theories and examples through six academic fields such as economics, geography, regional development, business administration, sociology and pedagogy and their educational back-ground. By means of establishing the incubation stage in the development of cluster, core elements are composed in the stages of birth, incubation and evolution in subsequent manner. A promoting strategy will be suggested through the implication of core elements in the reestablished stages.

• Genomics & Informatics
• /
• 제16권4호
• /
• pp.23.1-23.5
• /
• 2018

Virus taxonomy was initially determined by clinical experiments based on phenotype. However, with the development of sequence analysis methods, genotype-based classification was also applied. With the development of genome sequence analysis technology, there is an increasing demand for virus taxonomy to be extended from in vivo and in vitro to in silico. In this study, we verified the consistency of the current International Committee on Taxonomy of Viruses taxonomy using an in silico approach, aiming to identify the specific sequence for each virus. We applied this approach to norovirus in Caliciviridae, which causes 90% of gastroenteritis cases worldwide. First, based on the dogma "protein structure determines its function," we hypothesized that the specific sequence can be identified by the specific structure. Firstly, we extracted the coding region (CDS). Secondly, the CDS protein sequences of each genus were annotated by the conserved domain database (CDD) search. Finally, the conserved domains of each genus in Caliciviridae are classified by RPS-BLAST with CDD. The analysis result is that Caliciviridae has sequences including RNA helicase in common. In case of Norovirus, Calicivirus coat protein C terminal and viral polyprotein N-terminal appears as a specific domain in Caliciviridae. It does not include in the other genera in Caliciviridae. If this method is utilized to detect specific conserved domains, it can be used as classification keywords based on protein functional structure. After determining the specific protein domains, the specific protein domain sequences would be converted to gene sequences. This sequences would be re-used one of viral bio-marks.

• 한국컴퓨터정보학회논문지
• /
• 제17권7호
• /
• pp.67-76
• /
• 2012

본 논문에서는 보행자에 장착된 카메라로부터 입력된 자연영상에서 외부 환경 변화에 강인한 적응적인 보도 경계선 검출 기법을 제안한다. 제안하는 방법은 두 단계로 구성되어 안정적으로 보도 영역을 분할한다. 첫 번째 단계에서는 복잡한 외부 환경에 강인하도록 적응적인 임계치를 이용하는 에지 검출 방법을 통해 소실선과 소실점을 추출한다. 두 번째 단계에서는 소실점에 기반하는 VRay를 이용한 방사형 영역 분할법을 통해 보도의 경계선을 검출한다. 성능평가를 위해서는제안된 에지 검출 방법과 케니 에지 검출기와의 비교를통해 제안된 방법이 외부 환경의 조명조건 변화에 강건함을 확인하였으며, VRay의 가상 광선의 길이 변화에 따른 영역분할 결과를 비교하여 방사형 영역 분할법의 타당성을 입증하였다.

• Journal of Forest and Environmental Science
• /
• 제39권4호
• /
• pp.213-224
• /
• 2023

As abnormal weather events due to climate change continue to rise, landslide damage is also increasing. Given the substantial time and financial resources required for post-landslide recovery, it becomes imperative to formulate a proactive response plan. In this regard, landslide susceptibility analysis has emerged as a valuable tool for establishing preemptive measures against landslides. Accordingly, this study conducted an annual landslide susceptibility analysis using the history of landslides that occurred over many years in the Jeolla region, and analyzed areas with a high potential for landslides in the Jeolla region. The analysis employed an ensemble model that amalgamated 10 data-based models, aiming to mitigate uncertainties associated with a single-model approach. Furthermore, based on the cumulative data regarding landslide susceptible areas, this research identified regions vulnerable to recurring landslide damage in Jeolla region and proposed specific strategies for utilizing this information at various levels, including local government initiatives, adaptation plan development, and development approval processes. In particular, this study outlined approaches for local government utilization, the determination of adaptation plan types, and considerations for development permits. It is anticipated that this research will serve as a valuable opportunity to underscore the significance of information concerning regions vulnerable to recurring landslide damage.

• Journal of Information Processing Systems
• /
• 제15권1호
• /
• pp.67-90
• /
• 2019

Accurate detection, tracking and analysis of human movement using robots and other visual surveillance systems is still a challenge. Efforts are on to make the system robust against constraints such as variation in shape, size, pose and occlusion. Traditional methods of detection used the sliding window approach which involved scanning of various sizes of windows across an image. This paper concentrates on employing a state-of-the-art, hierarchical graph based method for segmentation. It has two stages: part level segmentation for color-consistent segments and object level segmentation for category-consistent regions. The tracking phase is achieved by employing SIFT keypoint descriptor based technique in a combined matching and tracking scheme with validation phase. Localization of human region in each frame is performed by keypoints by casting votes for the center of the human detected region. As it is difficult to avoid incorrect keypoints, a consensus-based framework is used to detect voting behavior. The designed methodology is tested on the video sequences having 3 to 4 persons.