• 한국컴퓨터정보학회:학술대회논문집
• /
• 한국컴퓨터정보학회 2011년도 제43차 동계학술발표논문집 19권1호
• /
• pp.55-58
• /
• 2011

This paper proposes a method of occluded number recognition by matching interest points. Interest points of input pattern are found via SURF features extracting and matched to interest points of clusters in database following three steps: SURF matching, coordinate matching and SURF matching on coordinate matched points. Then the satisfied interest points are counted to compute matching rate of each cluster. The input pattern will be assigned to cluster having highest matching rate. We have experimented our method to different numerical fonts and got encouraging results.

• 한국측량학회지
• /
• 제13권2호
• /
• pp.299-307
• /
• 1995

본 논문은 수치영상으로부터 Digital Photogrammetry와 Computer Vision 분야에서 위치결정 및 3차원 정보의 자동추출을 위한 기본단계인 Distinct Point 추출기법중 F rsner interest operate에 관한 연구이다. Gradien에 기초한 Forstner interest operator는 Orientation-invariant의 특징을 가지고 있으며 소정의 Subpixel정확도를 얻을 수 있다. 본 연구에서는 Fostner interest operator에서 얻어진 Comer Points와 Circular Features를 구분하기 위한 방법으로 F-test를 적용하였으며 Nosie가 Forstner interest operator에 미치는 영향을 고찰하였고 실제 사진영상에 Forstner interest operator를 도입하여 실효성에 바탕을 둔 적용 여부를 검증하였다.

• 한국측량학회지
• /
• 제28권3호
• /
• pp.353-359
• /
• 2010

본 논문은 Scale Invariant Feature Transform(SIFT) 기술자를 이용한 매칭 방법을 개선하여 고해상도영상에서 보다 많은 매칭쌍(tie points)을 추출함으로써 고해상도영상 자동기하보정의 결과향상을 목적으로 한다. 이를 위해 기준(reference)영상과 대상(sensed)영상의 특징점(interest points)간의 위치관계를 추가적으로 이용하여 매칭쌍을 추출하였다. SIFT 기술자를 이용하여 어핀(affine)변환계수를 추정한 후, 이를 통해 대상영상의 특징점 좌표를 기준영상 좌표체계로 변환하였다. 변환된 대상영상의 특징점과 기준영상의 특징점간의 공간거리(spatial distance)정보를 이용하여 최종적으로 매칭쌍을 추출하였다. 추출된 매칭쌍으로 piecewise linear function을 구성하여 고해상도 영상간 자동기하보정을 수행하였다. 제안한 기법을 통하여, 기존 SIFT 기법에 의해 추출한 결과에 비해 영상 전역에 걸쳐 고르게 분포된 다수의 매칭쌍을 추출할 수 있었다.

• 경영과학
• /
• 제18권1호
• /
• pp.119-133
• /
• 2001

It is important to find the equilibrium level of real interest rate for it affects real and financial sector of economy. However, it is difficult to find the equilibrium level because like the most macroeconomic model the real interest model has parameter instability problem caused by structural change and it is supported by various theories and definitions. Hence, in order to cover these problems structural change detection model of real interest rate is developed to combine the real interest rate equilibrium model and the procedure to detect structural change points. 3 equations are established to find various effects of other interest-related macroeconomic variables and from each equation, structural changes are found. Those structural change points are consistent with common expectation. Oil Crisis (December, 1987), the starting point of Economic Stabilization Policy (January, 1982), the starting point of capital liberalization (January, 1988), the starting and finishing points of Interest deregulation (January, 1992 and December, 1994), Foreign Exchange Crisis (December, 1977) are detected as important points. From the equation of fisher and real effects, real interest rate level is estimated as 4.09% (October, 1988) and dependent on the underlying model, it is estimated as 0%∼13.56% (October, 1988), so it varies so much. It is expected that this result is connected to the large scale simultaneous equations to detect the parameter instability in real time, so induces the flexible economic policies.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제8권5호
• /
• pp.1676-1689
• /
• 2014

At present, the Content-Based Image Retrieval (CBIR) system has become a hot research topic in the computer vision field. In the CBIR system, the accurate extractions of low-level features can reduce the gaps between high-level semantics and improve retrieval precision. This paper puts forward a new retrieval method aiming at the problems of high computational complexities and low precision of global feature extraction algorithms. The establishment of the new retrieval method is on the basis of the SIFT and Harris (APISH) algorithm, and the salient region of interest points (SRIP) algorithm to satisfy users' interests in the specific targets of images. In the first place, by using the IPDSH and SRIP algorithms, we tested stable interest points and found salient regions. The interest points in the salient region were named as salient interest points. Secondary, we extracted the pseudo-Zernike moments of the salient interest points' neighborhood as the feature vectors. Finally, we calculated the similarities between query and database images. Finally, We conducted this experiment based on the Caltech-101 database. By studying the experiment, the results have shown that this new retrieval method can decrease the interference of unstable interest points in the regions of non-interests and improve the ratios of accuracy and recall.

• 한국항행학회논문지
• /
• 제16권2호
• /
• pp.386-393
• /
• 2012

본 논문에서는 객체의 관심점(interest points)에 대한 지역 특징 기술자를 이용하여 이미지나 동영상에서 다수의 관심 객체를 효과적으로 인식하고 추적하기 위한 기법을 제안한다. 이를 위해 먼저 대상이 되는 객체를 포함하는 다양한 이미지를 수집하고 SURF 알고리즘을 적용하여 객체의 관심점과 그들에 대한 지역 특징 기술자를 생성한다. 지역 특징에 대한 통계적인 분석을 통하여 관심점들 중에서 해당 객체의 특성을 가장 잘 표현하는 대표점(representative points)을 선택하고 이를 바탕으로 이미지에 존재하는 객체를 인식한다. 또한, 지역 특징 기술자의 정합을 응용하여 각 SURF 지점들의 움직임 벡터를 생성하고 이를 기반으로 실시간으로 객체를 추적한다. 제안하는 기법은 모든 객체를 독립적으로 다루기 때문에, 여러 개의 객체를 동시에 인식하고 추적할 수 있다. 다양한 실험을 통해, 동영상에서 객체의 존재 여부 및 종류를 신속하게 판별하고 관심 객체의 추적을 효과적으로 수행할 수 있음을 보인다.

• Journal of Information Processing Systems
• /
• 제11권4호
• /
• pp.528-537
• /
• 2015

Text in images is one of the most important cues for understanding a scene. In this paper, we propose a novel approach based on interest points to localize text in natural scene images. The main ideas of this approach are as follows: first we used interest point detection techniques, which extract the corner points of characters and center points of edge connected components, to select candidate regions. Second, these candidate regions were verified by using tensor voting, which is capable of extracting perceptual structures from noisy data. Finally, area, orientation, and aspect ratio were used to filter out non-text regions. The proposed method was tested on the ICDAR 2003 dataset and images of wine labels. The experiment results show the validity of this approach.

• 한국측량학회:학술대회논문집
• /
• 한국측량학회 2004년도 춘계학술발표회논문집
• /
• pp.171-176
• /
• 2004

Image matching is fundamental process in photogrammetry and computer vision to identify and to measure corresponding features on the multiple images. Uniqueness of the matching entities and robustness of the algorithm are the key issues that have influence on quality of the matching result. The optimal solution could be obtained by utilizing appropriate matching entities in the first place. In this study, candidate matching points were extracted by interest operator, and an area-based matching method was applied with characteristics of the gray value distribution as the matching entities. The characteristic information is based on the concept of "intrinsic image" (or parameter image). The information was utilized as additional and/or complementary matching entities. Matching on interest points with the characteristic information resulted in high quality of matching because matching windows were created with surrounding pixels of the interest points that contain distinct and unique features. The experiment shows that matching quality and reliability increase by exploiting interest operator, and the characteristic information has potential to be matching entity.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2003년도 ICCAS
• /
• pp.673-676
• /
• 2003

This paper deals with the problem of moving object detection and location in computer vision. We describe a new object-dependent motion analysis method for tracking target in an image sequence taken from a moving platform. We tackle these tasks with three steps. First, we make an active contour model of a target in order to build some of low-energy points, which are called kernels. Then we detect interest points in two windows called tracking windows around a kernel respectively. At the third step, we decide the correspondence of those detected interest points between tracking windows by the probabilistic relaxation method In this algorithm, the detecting process is iterative and begins with the detection of all potential correspondence pair in consecutive image. Each pair of corresponding points is then iteratively recomputed to get a globally optimum set of pairwise correspondences.

• 방송공학회논문지
• /
• 제22권1호
• /
• pp.70-77
• /
• 2017

다양한 사람계수 측정 방법중에서 동적 관심점을 이용하는 지도-기반 기법은 우수한 성능을 보여준다. 그러나 정적인 사람의 계수측정은 정적 관심점이 배경에 포함되기 때문에 어려움이 있다. 계수에 정적인 사람을 포함하기 위해서 정적인 사람이 정적점과 배경을 구별하는 것이 필요하다. 본 논문에서는 동적 및 정적 점들을 고려하는 사람계수 방법을 제안한다. 제안방법은 모션정보를 활용하여 두 점을 분리한다. 그러면 정적인 사람의 정적점들은 전경 마스크 처리 및 점 패턴 분석를 하여 분류된다. 실험결과에서는 제안 방법이 정적인 사람을 계수에 포함하기 때문에 보다 정확한 사람계수 값을 얻는다. 또한 배경 갱신을 이용함으로써 배경 변화에 따른 정적점 오분류 문제를 해결한다.