• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제14권9호
• /
• pp.3762-3781
• /
• 2020

Infrared pedestrian target recognition is a vital problem of significant interest in computer vision. In this work, a novel infrared pedestrian target recognition method that uses heterogeneous features with multi-kernel transfer learning is proposed. Firstly, to exploit the characteristics of infrared pedestrian targets fully, a novel multi-scale monogenic filtering-based completed local binary pattern descriptor, referred to as MSMF-CLBP, is designed to extract the texture information, and then an improved histogram of oriented gradient-fisher vector descriptor, referred to as HOG-FV, is proposed to extract the shape information. Second, to enrich the semantic content of feature expression, these two heterogeneous features are integrated to get more complete representation for infrared pedestrian targets. Third, to overcome the defects, such as poor generalization, scarcity of tagged infrared samples, distributional and semantic deviations between the training and testing samples, of the state-of-the-art classifiers, an effective multi-kernel transfer learning classifier called MK-TrAdaBoost is designed. Experimental results show that the proposed method outperforms many state-of-the-art recognition approaches for infrared pedestrian targets.

• 한국생물정보학회:학술대회논문집
• /
• 한국생물정보시스템생물학회 2003년도 제2차 연례학술대회 발표논문집
• /
• pp.3-3
• /
• 2003

With biomedical literature expanding so rapidly, there is an urgent need to discover and organize knowledge extracted from texts. Although factual databases contain crucial information the overwhelming amount of new knowledge remains in textual form (e.g. MEDLINE). In addition, new terms are constantly coined as the relationships linking new genes, drugs, proteins etc. As the size of biomedical literature is expanding, more systems are applying a variety of methods to automate the process of knowledge acquisition and management. In my talk, I focus on the project, GENIA, of our group at the University of Tokyo, the objective of which is to construct an information extraction system of protein - protein interaction from abstracts of MEDLINE. The talk includes (1) Techniques we use fDr named entity recognition (1-a) SOHMM (Self-organized HMM) (1-b) Maximum Entropy Model (1-c) Lexicon-based Recognizer (2) Treatment of term variants and acronym finders (3) Event extraction using a full parser (4) Linguistic resources for text mining (GENIA corpus) (4-a) Semantic Tags (4-b) Structural Annotations (4-c) Co-reference tags (4-d) GENIA ontology I will also talk about possible extension of our work that links the findings of molecular biology with clinical findings, and claim that textual based or conceptual based biology would be a viable alternative to system biology that tends to emphasize the role of simulation models in bioinformatics.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제17권12호
• /
• pp.3383-3397
• /
• 2023

Scene graphs serve as semantic abstractions of images and play a crucial role in enhancing visual comprehension and reasoning. However, the performance of Scene Graph Generation is often compromised when working with biased data in real-world situations. While many existing systems focus on a single stage of learning for both feature extraction and classification, some employ Class-Balancing strategies, such as Re-weighting, Data Resampling, and Transfer Learning from head to tail. In this paper, we propose a novel approach that decouples the feature extraction and classification phases of the scene graph generation process. For feature extraction, we leverage a transformer-based architecture and design an adaptive calibration function specifically for predicate classification. This function enables us to dynamically adjust the classification scores for each predicate category. Additionally, we introduce a Distribution Alignment technique that effectively balances the class distribution after the feature extraction phase reaches a stable state, thereby facilitating the retraining of the classification head. Importantly, our Distribution Alignment strategy is model-independent and does not require additional supervision, making it applicable to a wide range of SGG models. Using the scene graph diagnostic toolkit on Visual Genome and several popular models, we achieved significant improvements over the previous state-of-the-art methods with our model. Compared to the TDE model, our model improved mR@100 by 70.5% for PredCls, by 84.0% for SGCls, and by 97.6% for SGDet tasks.

• 한국공간정보시스템학회 논문지
• /
• 제11권2호
• /
• pp.39-44
• /
• 2009

In order to make the effect of the navigation system more direct, the paper proposes a thought of vehicle navigation system based on Video-GIS. A semantic framework has been defined whose core is focused on the integration and interaction of video and spatial information, which supports full content retrieval based on multimodal metadata extraction and fusion, and supports kinds of wireless access mode. Furthermore, requirements of prototype system are discussed. Then the design and implementation of framework are discussed. Next, describe the key ideas and technologies involved. Finally, we point out its future research trend.

• 대한원격탐사학회지
• /
• 제38권6_4호
• /
• pp.1889-1900
• /
• 2022

도로는 현대사회의 기능이 물리적으로 작동하는 데 필수불가결한 요소이다. 교통상황정보에 비해 갱신 주기가 긴 도로공간 정보를 더 빠르고 정확하게 생성할 필요가 있다. 본 연구에서는 그 방법의 일환으로 아리랑 3호와 아리랑 3A호의 위성영상에 pan-sharpening 영상융합 기법을 적용하여 공간해상도를 향상시킨 영상자료를 최근 활발히 연구가 진행되고 있는 semantic segmentation 기법을 활용한 도로 추출에 활용하고자 하였다. 확보한 영상은 U-Net 기반의 segmentation 기법에 매사추세츠 도로데이터와 함께 투입하여 훈련하였고 아리랑 위성 융합영상의 모델 적용 가능성을 평가하였다. 훈련 및 검증 결과, 모델에 투입하는 영상에 대해 일정한 조건이 유지되는 한 일정한 모델 예측 성능을 유지하는 것으로 나타났다. 따라서 그림자와 지표면 상태와 같은 모델에 영향을 미치는 주변 환경 조건의 영향을 최소화하는 방법을 적용하여 풍부한 훈련자료를 구성한다면 아리랑위성과 같은 위성 영상의 활용 가능성이 더욱 높아질 것으로 기대된다.

• International Journal of Computer Science & Network Security
• /
• 제22권10호
• /
• pp.406-412
• /
• 2022

Scene classification of very high-resolution (VHR) imagery can attribute semantics to land cover in a variety of domains. Real-world application requirements have not been addressed by conventional techniques for remote sensing image classification. Recent research has demonstrated that deep convolutional neural networks (CNNs) are effective at extracting features due to their strong feature extraction capabilities. In order to improve classification performance, these approaches rely primarily on semantic information. Since the abstract and global semantic information makes it difficult for the network to correctly classify scene images with similar structures and high interclass similarity, it achieves a low classification accuracy. We propose a VHR remote sensing image classification model that uses extracts the global feature from the original VHR image using an EfficientNet-V2L CNN pre-trained to detect similar classes. The image is then classified using a multilayer perceptron (MLP). This method was evaluated using two benchmark remote sensing datasets: the 21-class UC Merced, and the 38-class PatternNet. As compared to other state-of-the-art models, the proposed model significantly improves performance.

• 한국정보통신학회:학술대회논문집
• /
• 한국해양정보통신학회 2008년도 추계종합학술대회 B
• /
• pp.532-535
• /
• 2008

디지털 콘텐츠의 대부분을 차지하는 동영상에 대한 검색 서비스가 필수 기능으로 대두되고 있으며, 검색 서비스를 수행하는 시스템은 최신 기술을 접목시켜 보다 지능적이고, 의미적인 검색을 할 수 있는 검색 엔진이나 지능형 검색 기법 등의 필요성이 점차 증대되고 있다. 이에 본 논문에서는 디지털 콘텐츠 데이터에 대한 특성요소 분석 및 검색 기술과 구현, 감성어휘기반 분석 및 검색 방안을 위해 멀티미디어 콘텐츠 데이터의 구조 설계와 분석 관리 도구 및 의미론적 특성요소 추출기술과 콘텐츠 내 컬러 정보 기반 감성처리 알고리즘을 제안하였다.

• 한국언어정보학회:학술대회논문집
• /
• 한국언어정보학회 2007년도 정기학술대회
• /
• pp.508-514
• /
• 2007

In order to extract some important information of a person from text, an extracting model was proposed. The person's name is recognized based on the maximal entropy statistic model and the training corpus. The sentences surrounding the person's name are analyzed according to the conceptual knowledge base. The three main elements of events, domain, situation and background, are also extracted from the sentences to construct the structure of events about the person.

• 한국정보과학회논문지:소프트웨어및응용
• /
• 제36권1호
• /
• pp.70-81
• /
• 2009

온톨로지는 관심 영역의 개념과 개념관계를 명시적으로 명세한 것을 말하며, 지식 표현의 대표적인 방법으로 인식되어 의미에 기반을 둔 정보의 추출, 지식 관리, 정보 공유 등 다양한 분야에서 온톨로지를 적용한 연구가 이루어지고 있다. 정보기술(IT) 기반의 경제/산업 분야에서 기업 간의 상호 협력을 위한 정보 공유 및 통합 연구에 온톨로지의 적용이 이루어지고 있다. 여러 업체가 물류 주체로 참여하며 물품의 이동, 보관, 배송 등을 계획하고 관리하는 물류 분야에서도 원활한 공급체인관리나 물류관리를 위한 물류정보의 통합과 정보공유 연구가 많이 이루어지고 있다. 최근에는 물품마다 부여한 고유의 식별코드에 의한 물품의 추적과 관리, 물류 과정의 가시성 제공 둥의 요구가 발생하면서 물류 과정에 흩어져 있는 물류정보의 통합 제공 요구가 증가하고 있다. 이에 본 논문에서는 물류 과정에서 발생하는 데이타를 의미에 기반을 두고 해석하고 통합하기 위한 지식자원으로 물류 도메인 온톨로지를 제시한다. 물품을 식별하는 고유 식별코드인 전자물품코드(EPC)로 물품의 추적과 관리가 이루어지는 EPC 네트워크 기반의 물류 환경에서 발생하는 EPC 이벤트 데이타를 의미에 따라 표현하고 이벤트 데이타의 내포된 의미를 해석할 수 있는 개념과 개념관계를 표현하는 데 초점을 맞추어 온톨로지를 구성하였다. 그리고 EPC 네트워크 기반의 물류 환경에서 물품의 위치, 상태, 이동경로 등 물류 관리를 위한 정보추출 과정에서 물류 도베인 온톨로지가 이용될 수 있음을 물류 시나리오를 통해 보였다.

• Journal of Information Technology Applications and Management
• /
• 제26권1호
• /
• pp.65-75
• /
• 2019

As the use of semantic web based on XML increases in the field of data management, a lot of studies to extract useful information from the data stored in ontology have been tried based on association rule mining. Ontology data is advantageous in that data can be freely expressed because it has a flexible and scalable structure unlike a conventional database having a predefined structure. On the contrary, it is difficult to find frequent patterns in a uniformized analysis method. The goal of this study is to provide a basis for extracting useful knowledge from ontology by searching for frequently occurring subgraph patterns by applying transaction-based graph mining techniques to ontology schema graph data and instance graph data constituting ontology. In order to overcome the structural limitations of the existing ontology mining, the frequent pattern search methodology in this study uses the methodology used in graph mining to apply the frequent pattern in the graph data structure to the ontology by applying iterative node chunking method. Our suggested methodology will play an important role in knowledge extraction.