• IEIE Transactions on Smart Processing and Computing
• /
• 제3권6호
• /
• pp.366-371
• /
• 2014

This paper reports a machine learning approach for image object detection. Object detection and localization in a wild image, such as a STL-10 image dataset, is very difficult to implement using the traditional computer vision method. A convolutional neural network is a good approach for such wild image object detection. This paper presents an object detection application using a convolutional neural network with pretrained feature vector. This is a very simple and well organized hierarchical object abstraction model.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권10호
• /
• pp.3729-3749
• /
• 2021

At present, deep convolution network-based salient object detection (SOD) has achieved impressive performance. However, it is still a challenging problem to make full use of the multi-scale information of the extracted features and which appropriate feature fusion method is adopted to process feature mapping. In this paper, we propose a new adjacency auxiliary network (AANet) based on multi-scale feature fusion for SOD. Firstly, we design the parallel connection feature enhancement module (PFEM) for each layer of feature extraction, which improves the feature density by connecting different dilated convolution branches in parallel, and add channel attention flow to fully extract the context information of features. Then the adjacent layer features with close degree of abstraction but different characteristic properties are fused through the adjacent auxiliary module (AAM) to eliminate the ambiguity and noise of the features. Besides, in order to refine the features effectively to get more accurate object boundaries, we design adjacency decoder (AAM_D) based on adjacency auxiliary module (AAM), which concatenates the features of adjacent layers, extracts their spatial attention, and then combines them with the output of AAM. The outputs of AAM_D features with semantic information and spatial detail obtained from each feature are used as salient prediction maps for multi-level feature joint supervising. Experiment results on six benchmark SOD datasets demonstrate that the proposed method outperforms similar previous methods.

• 한국전자거래학회지
• /
• 제18권2호
• /
• pp.95-110
• /
• 2013

본 논문은 기계학습 기법에 기반한 웹문서 자동분류 시스템의 성능을 높이기 위한 새로운 형태의 특징가공 기법을 제안한다. 제안 기법은 하이퍼텍스트 웹문서에 대한 자동분류를 효과적으로 수행하기 위해 하이퍼링크 관계를 활용하여 특징 집합을 확장시킨다. 웹문서는 하이퍼링크 관계를 통해 서로 연결된 구조를 가지며, 그 관계는 많은 경우 연관도가 높은 문서들 간에 존재한다. 이러한 링크 정보가 분류모델의 주요 인자가 되는 특징 집합의 질을 높이는 중요한 역할을 수행할 수 있다. 제안 기법의 기본 아이디어는 워드넷 온톨로지를 기반으로 분류 대상 문서와 인접 문서들에 포함된 단어(특징)들 간의 의미적 유사도를 평가함으로써 다수의 특징들로 구성된 추상화된 개념적 특징을 생성하는 것이다. 여기서 유사도 함수는 워드넷 안에서 특징들 간의 상/하위어 관계 정보를 정량적으로 계산하게 된다. 분류모델의 구축시 추상화된 개념 특징은 일반 특징과 동일하게 간주하여 보다 정확한 분류 모델을 구축하는데 기여한다. Web-KB 문서집합을 이용한 실험을 통해 제안 기법이 기존 기법 보다 우수함을 보였다.

• 한국의상디자인학회지
• /
• 제25권3호
• /
• pp.97-111
• /
• 2023

Dynamic orphism images expressed with abstraction and construction in color are often found in modern fashion. Orphism stressing color is one of the many art trends that influence fashion. This work aims to examin its expressive effects and characteristics that appear in fashion and provide basic materials for fashion design study. This study proceeds with the following steps. First, the study takes on a theoretical examination of the orphism trend using a literature review. Second, based on its results, modern fashion's expressive features influenced by orphism are analyzed. Third, the study draws expression effects in modern fashion from these characteristics. The study coverage is confined to domestic and foreign collections released in the 21st century. The study contents come from an Internet-base domestic and international database and published material, including dissertations and books. Orphism expression effects are found in the following ways in modern fashion. First, it lies in the abstraction effect of color. This feature arrays flat geometric figure on the clothing surface and applies a strong sense of color inside it, as if color has an abstractive structured design. Second, it gives a dynamic sense visually to the clothes all at once by stressing the relation among and between the colors with powerful contrast. Third, it has a collage effect of color. This effect develops into an organic combination composed of geometric plane figures with intense complementary colors in a single piece of clothing. Expression types that presents the orphism effect in fashion are divided into clothing showing traditional orphism characteristics faithfully and clothing expressing conventional orphism characteristics mixed with modern trends. Further developed orphism fashion can be classified by the playful type, with graffiti added onto an orphism color structure, and the destructive type, where traditional orphism color orders are taken apart and reformed into a new color order.

• 한국정보통신학회논문지
• /
• 제3권2호
• /
• pp.373-382
• /
• 1999

상속성과 병행성은 병행 객체지향 언어에서 가장 주된 개념이며 특히 코드의 재사용에 있어서 매우 중요하다. 이들은 객체의 병행 수행을 통해 최대의 계산력과 모델링 능력을 제공한다. 하지만, 병행 객체와 상속성은 서로 상충되는 특성을 가지고 있으며, 이들의 동시 사용은 병행 객체들의 무결성을 유지하기 위해 상속된 메소드들의 코드 재정의를 요구하는 상속 변칙 문제를 발생시킨다. 본 논문에서는 상속 변칙의 해결을 위해 캡슐화된 객체의 내부 상태들이 객체의 외부 인터페이스의 한 부분으로 이용될 수 있는 상태 추상화 개념을 도입하였다. 또한, 메소드들을 효과적으로 상속할 수 있는 상속 인터페이스 메커니즘을 설계하였으며 이를 통해 전형적인 상속 변칙 문제를 해결하였다.

• 한국자동차공학회논문집
• /
• 제1권3호
• /
• pp.83-94
• /
• 1993

Design and production of wiring for automobile systems provide the background for the development of concurrent engineering framwork. Key issues include with different parts of the wiring design process and the development of hierarchical representations that capture the different characteristics(e.g., connectivity, geometry) of the harnesses. The abstraction of design information results in features, while the abstraction of operations leads to the development of agents. These abstracts are essential for efficient transactions among people and computer tools in a domain that involves numerous interacting constraints. We have decomposed the wiring design process into four basic categories of tasks that are each addressed by an "agent" : the Environment Editor(EE), the Free Space Manager(FSM), the Wiring Editor(WE), and the Part Selector(PS). In this paper the strategy for the problem decomposition, the definition of features, and the ways in which features are used by various agents, are discussed. We conclude with a discussion of some of the issues raised by the project and the steps underway to address them.

• 한국지능시스템학회:학술대회논문집
• /
• 한국지능시스템학회 2008년도 춘계학술대회 학술발표회 논문집
• /
• pp.432-434
• /
• 2008

본 논문은 필터링 수행율을 향상시키기 위해, DTW 알고리즘을 제시한다. 여기에서 말하는 컨텐츠 필터링은 음성의 특징을 사용해서 컨텐츠를 구분하는 것을 확인하는 기술이다. 즉, 이 방법이 일반 컨텐츠와 성인 컨텐츠를 구분한다. 음성에 대한 정보를 추출하는 방법이 컨텐츠를 필터링하는데 있어서 기여를 할 수 있다. 즉, DTW 알고리즘을 사용하여 필터링 인식률을 향상하는 방법이라고 제안을 한다. 마지막으로, 본 논문에서 제안한 방법의 적용 가능성과 일반성을 평가하기위하여 수치적인 예를 적용한다. 제안하는 성질의 정확도를 시험하기 위해서 실험을 제공하였다. 결과적으로 일반 컨텐츠와 성인 컨텐츠 특성의 차이를 알았다. 추후에 이 성질을 필터링 성능 향상에 응용할 수 있다.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2005년도 학술대회 논문집 정보 및 제어부문
• /
• pp.625-627
• /
• 2005

Existent distance measurement that use camera is method that use both Stereo Camera and Monocular Camera, There is shortcoming that method that use Stereo Camera is sensitive in effect of a lot of expenses and environment variables, and method that use Monocular Camera are big computational complexity and error. In this study, reduce expense and error using Monocular Camera and I suggest algorithm that measure distance, Extract features using scale Invariant features Transform(SIFT) for distance measurement, and this measures distance through features matching and geometrical analysis, Proposed method proves measuring distance with wall by geometrical analysis free wall through feature point abstraction and matching.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2002년도 ICCAS
• /
• pp.34.5-34
• /
• 2002

In this paper we propose a new method for modeling indoor environment with stereo camera and suggest a localization method for mobile robot navigation on the basis of it. From the viewpoint of easiness in map building and exclusion of artificiality, the main idea of this paper is that environment is represented as global topological map and each node has omni-directional metric and color information by using stereo camera and pan/tilt mechanism. We use the depth and color information itself in image pixel as feature for environmental abstraction. In addition, we use only the depth and color information at horizontal centerline in image, where optical axis is passing. The usefulness of this m...

• ETRI Journal
• /
• 제29권4호
• /
• pp.507-515
• /
• 2007

Skeleton transform of which the medial axis transform is the most popular has been proposed as a useful shape abstraction tool for the representation and modeling of human posture. This paper explains this proposition with a description of the areas in which skeletons could serve to enable the representation of shapes. We present algorithms for two-dimensional posture modeling using the developed simplified shock graph (SSG). The efficacy of SSG extracted feature vectors as shape descriptors are also evaluated using three different classifiers, namely, decision tree, multilayer perceptron, and support vector machine. The paper concludes with a discussion of the issues involved in using shock graphs to model and classify human postures.