• Journal of Korea Multimedia Society
• /
• v.6 no.4
• /
• pp.746-752
• /
• 2003

In this paper, we proposed an image retrieval system with a simple and effective relevance feedback, called RAP(Reward and Punishment) algorithm. First, color and texture features were extracted from the images. Next, the extracted feature values were used for image retrieval in various forms. We applied the relevance feedback to the initial retrieved images from the image retrieval system, and compared its result with that of the conventional system. In the experiment using the test image database of 16 class 512 images, the proposed system showed the better retrieval performance of about 10∼l7 % than that of the conventional INRIA system in each relevance feedback step.

• Proceedings of the KIEE Conference
• /
• 1987.07b
• /
• pp.1077-1080
• /
• 1987

Images are 2-dimensional projection of 3-dimensional scenes and many problems of scene analysis arise due to inherent depth ambiguities in a monocular 2-D image. Therefore, depth recovery is a crucial problem in image understanding. This paper proposes modified algorithm which is focused on accurate correspondnce in stereo vision. The feature we use is zero-crossing points and the similarity measure with two property evaluation function is used to estimate initial probability. And we introduce relaxation technique for accurate and global correspondence.

• Annual Conference on Human and Language Technology
• /
• 2018.10a
• /
• pp.347-351
• /
• 2018

이 논문은 대화 시스템에서 질의를 이해하기 위해 딥 러닝 모델을 통해 추출된 자동 추출 자질을 이용하여 문장의 유사성을 분석하는 방법에 대해 기술한다. 문장 간 유사성을 분석하기 위한 자동 추출 자질로써, 문장 내 표현 순차적 정보를 반영하기 위한 RNN을 이용하여 생성한 문장 벡터와, 어순에 관계 없이 언어 모델을 학습하기 위한 CNN을 이용하여 생성한 문장 벡터를 사용한다. 이렇게 자동으로 추출된 문장 임베딩 자질은 금융서비스 대화에서 입력 문장을 분류하거나 문장 간 유사성을 분석하는데 이용된다. 유사성 분석 결과는 질의 문장과 관련된 FAQ 문장을 찾거나 답변 지식을 찾는데 활용된다.

• Journal of the Semiconductor & Display Technology
• /
• v.5 no.2 s.15
• /
• pp.27-31
• /
• 2006

As more and more digital images are made by various applications, image retrieval becomes a primary concern in technology of multimedia. This paper presents color based descriptor that uses information of color distribution in color images which is the most basic element for image search and performance of proposed visual feature is evaluated through the simulation. In designing the image search descriptor used color histogram, HSV, Daubechies 9/7 and 2 level wavelet decomposition provide better results than other parameters in terms of computational time and performances. Also histogram quadratic matrix outperforms the sum of absolute difference in similarity measurements, but spends more than 60 computational times.

• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.2393-2396
• /
• 2005

We present a method to detect on-road vehicle using geometric invariant of feature points on side planes of the vehicle. The vehicles are assumed into a set of planes and the invariant from motion information of features on the plane segments the plane from the theory that a geometric invariant value defined by five points on a plane is preserved under a projective transform. Harris corners as a salient image point are used to give motion information with the normalized correlation centered at these points. We define a probabilistic criterion to test the similarity of invariant values between sequential frames. Experimental results using images of real road scenes are presented.

• Journal of the Korea Society of Computer and Information
• /
• v.7 no.4
• /
• pp.83-88
• /
• 2002

Large sets of conceptual schema have been constructed for database design. In recent times, the need of analytic aid for schema reuse is increasing. In this paper, it is presented analysis technique of conceptual schema, and experimented schema analysis for extraction of reference schema. It is desirable for integration of related schema to have been applied in case of similarity value above 0.6. Reference schema which is created through the analysis technique enable to describe concepts of them and can be the way of schema reuse. And a feature analysis can be effective measure to set details of analytic data which is necessary for extraction of reference schema.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.11 no.3
• /
• pp.1650-1669
• /
• 2017

In this study, each pixel in an ear is used as a centroid to generate a cake. Subsequently the major axis length of this cake is computed and obtained. This obtained major axis length serves as a feature to recognize an ear. Later, the ear hole is used as a centroid and a 16-circle template is generated to extract the major axis lengths of the ear. The 16-circle template extracted signals are used to recognize an ear. In the next step, a ring-to-line mapping technique is used to map these major axis lengths to several straight-line signals. Next, the complex plane vector computing technique is used to determine the similarity of these major axis lengths, whereby a solution to the image-rotating problem is achieved. The aforementioned extracted signals are also compared to the ones that are extracted from its neighboring pixels, whereby solving the image-shifting problem. The algorithm developed in this study can precisely identify an ear image by solving the image rotation and image shifting problems.

• 제어로봇시스템학회:학술대회논문집
• /
• 2001.10a
• /
• pp.101.1-101
• /
• 2001

This paper describes a grid mapping for a vision based mobile robot in uncertain indoor environment. The map building is a prerequisite for navigation of a mobile robot and the problem of feature correspondence across two images is well known to be of crucial Importance for vision-based mapping We use a stereo matching algorithm obtained by singular value decomposition of an appropriate correspondence strength matrix. This new correspondence strength means a correlation weight for some local measurements to quantify similarity between features. The visual range data from the reconstructed disparity image form an occupancy grid representation. The occupancy map is a grid-based map in which each cell has some value indicating the probability at that location ...

• International Journal of CAD/CAM
• /
• v.6 no.1
• /
• pp.125-137
• /
• 2006

Nowadays, many commercial CAD systems support history-based, constraint-based and feature-based modeling. Unfortunately, most systems fail during the re-evaluation phase when various kind of topological changes occur. This issue is known as "persistent naming" which refers to the problem of identifying entities in an initial parametric model and matching them in the re-evaluated model. Most works in this domain focus on the persistent naming of atomic entities such as vertices, edges or faces. But very few of them consider the persistent naming of aggregates like shells (any set of faces). We propose in this paper a complete framework for identifying and matching any kind of entities based on their underlying topology, and particularly shells. The identifying method is based on the invariant structure of each class of form features (a hierarchical structure of shells) and on its topological evolution (an historical structure of faces). The matching method compares the initial and the re-evaluated topological histories, and computes two measures of topological similarity between any couple of entities occurring in both models. The naming and matching method has been implemented and integrated in a prototype of commercial CAD Software (Topsolid).

• Proceedings of the IEEK Conference
• /
• 1999.11a
• /
• pp.601-604
• /
• 1999

This Paper proposes the simple and efficient image retrieval algorithm using subregional texture features. In order to retrieve images in terms of its contents, it is required to obtain a precise segmentation. However, it is very difficult and takes a long computing time. Therefore. this paper proposes a simple segmentation method, which is to divide an image into high and low entropy regions by using Picture Information Measure (PIM). Also, in order to describe texture characteristics of each region, this paper suggest six different texture features produced on the basis of co-occurrence matrix. For an image retrieval system, a normalised correlation is adopted as a similarity function, which is not dependent on the range of each texture feature values. Finally, this proposed algorithm is applied to a various images and produces competitive results.