• 한국컴퓨터정보학회논문지
• /
• 제12권6호
• /
• pp.147-152
• /
• 2007

디지털 영상에서의 얼굴탐색은 얼굴인식을 위한 기본 단계이면서 인식 성능에 큰 영향을 미치는 중요한 처리 단계이다. 템플릿 정합 방식의 객체 검출방식에서 사용되어 얼굴 인식 등에서 좋은 성능을 보이는 Hausdorff 거리는 주어진 점의 집합들 사이에서 기하학적 유사도만을 고려한 측도이므로 원래의 영상이 포함하고 있는 다른 정보들을 추가적으로 이용함으로 효율을 높일 수 있다. 이러한 점에 착안하여 본 논문에서는 점들 사이에 서로 다른 정도를 측정하기 위해서 거리뿐만 아니라 점들 주위의 국지적 계조패턴 정보까지 포함하는 측도를 정의함으로써 보다 정밀한 템플릿 정합결과를 얻는 방법을 제안한다.

• 한국지능시스템학회논문지
• /
• 제14권7호
• /
• pp.816-820
• /
• 2004

클러스터링은 데이터 집합을 유사한 데이터 개체들의 클러스터들로 분할하여 데이터 속에 존재하는 의미 있는 정보를 얻는 과정이다. 클러스터링의 주요 쟁점은 고차원 데이터를 효율적으로 클러스터링하는 것과 최적화 문제를 해결하는 것이다. 본 논문에서는 SVM(Support Vector Machines)기반의 새로운 유사도 측정법과 효율적으로 클러스터의 개수를 생성하는 방법을 제안한다. 고차원의 데이터는 커널 함수를 이용해 Feature Space로 매핑시킨 후 이웃하는 클러스터와의 유사도를 측정한다. 이미 생성된 클러스터들은 측정된 유사도 값과 Δd 임계값에 의해서 원하는 클러스터의 개수를 얻을 수 있다. 제안된 방법을 검증하기 위하여 6개의 UCI Machine Learning Repository의 데이터를 사용한 결과, 제시된 클러스터의 개수와 기존의 연구와 비교하여 향상된 응집도를 얻을 수 있었다.

• 한국측량학회지
• /
• 제30권6_2호
• /
• pp.665-672
• /
• 2012

Amid rapidly increasing imagery inputs and their volume in a remote sensing imagery database, Content-Based Image Retrieval (CBIR) is an effective tool to search for an image feature or image content of interest a user wants to retrieve. It seeks to capture salient features from a 'query' image, and then to locate other instances of image region having similar features elsewhere in the image database. For a CBIR approach that uses texture as a primary feature primitive, designing a texture descriptor to better represent image contents is a key to improve CBIR results. For this purpose, an extended feature vector combining the Gabor filter and co-occurrence histogram method is suggested and evaluated for quantitywise and qualitywise retrieval performance criterion. For the better CBIR performance, assessing similarity between high dimensional feature vectors is also a challenging issue. Therefore a number of distance metrics (i.e. L1 and L2 norm) is tried to measure closeness between two feature vectors, and its impact on retrieval result is analyzed. In this paper, experimental results are presented with several CBIR samples. The current results show that 1) the overall retrieval quantity and quality is improved by combining two types of feature vectors, 2) some feature is better retrieved by a specific feature vector, and 3) retrieval result quality (i.e. ranking of retrieved image tiles) is sensitive to an adopted similarity metric when the extended feature vector is employed.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제13권3호
• /
• pp.1583-1598
• /
• 2019

Recently, software has increased in complexity and been applied in various industrial fields. As a result, the presence of software bugs cannot be avoided. Various bug severity prediction methodologies have been proposed, but their performance needs to be further improved. In this study, we propose a novel technique for bug severity prediction in cross projects such as Eclipse, Mozilla, WireShark, and Xamarin by using topic modeling and similarity (i.e., KL-divergence). First, we construct topic models from bug repositories in cross projects using Latent Dirichlet Allocation (LDA). Then, we find topics in each project that contain the most numerous similar bug reports by using a new bug report. Next, we extract the bug reports belonging to the selected topics and input them to a Naïve Bayes Multinomial (NBM) algorithm. Finally, we predict the bug severity in the new bug report. In order to evaluate the performance of our approach and to verify the difference between cross projects and single project, we compare it with the Naïve Bayes Multinomial approach; the Lamkanfi methodology, which is a well-known bug severity prediction approach; and an emotional similarity-based bug severity prediction approach. Our approach exhibits a better performance than the compared methods.

• 대한한의학원전학회지
• /
• 제35권4호
• /
• pp.1-9
• /
• 2022

Objectives : This study aims to increase efficiency of the prescription similarity analysis method that uses drug composition ratio. Methods : The controlled experiment compared result generation time, generated data quantity, and accuracy of results between previous and new analysis method on the 12,598 formulas and 61 prescription groups. Results : The control group took 346 seconds on average and generated 768,478 results, while the test group took 24 seconds and generated 241,739 results. The test group adopted a selective calculation method that only used overlapping data between two formulas instead of analyzing all number of cases. It simplified the data processing process, reducing the quantity of data that is required to be processed, leading to better system speed, as fast as 14.47 times more than previous analysis method with equal results. Conclusions : Efficiency for similarity analysis could be improved by reducing data span and simplifying the calculation processes.

• Nuclear Engineering and Technology
• /
• 제55권12호
• /
• pp.4591-4596
• /
• 2023

Single-photon emission computed tomography SPECT image reconstruction methods have a significant influence on image quality, with filtered back projection (FBP) and ordered subset expectation maximization (OSEM) being the most commonly used methods. In this study, we proposed newly-designed adaptive non-blind deconvolution with a structural similarity (SSIM) index that can take advantage of the FBP and OSEM image reconstruction methods. After acquiring brain SPECT images, the proposed image was obtained using an algorithm that applied the SSIM metric, defined by predicting the distribution and amount of blurring. As a result of the contrast to noise ratio (CNR) and coefficient of variation evaluation (COV), the resulting image of the proposed algorithm showed a similar trend in spatial resolution to that of FBP, while obtaining values similar to those of OSEM. In addition, we confirmed that the CNR and COV values of the proposed algorithm improved by approximately 1.69 and 1.59 times, respectively, compared with those of the algorithm involving an inappropriate deblurring process. To summarize, we proposed a new type of algorithm that combines the advantages of SPECT image reconstruction techniques and is expected to be applicable in various fields.

• International Journal of Internet, Broadcasting and Communication
• /
• 제12권1호
• /
• pp.111-115
• /
• 2020

In this paper, I proposed an effective technique for accurately predicting pixel values using edge components. Adjacent pixel values are similar to each other. That is, generally, similarity exists between adjacent pixels in an image. In the proposed algorithm, edge components are detected using the surrounding pixels in the first step, and pixel values are estimated using the edge components in the second step. Therefore, the prediction accuracy of the pixel value is improved and the prediction error is reduced. Pixel value prediction is a necessary technique for various applications such as image magnification and confidential data concealment. Experimental results show that the proposed method has higher prediction accuracy and fewer prediction error. Therefore, the proposed technique can be effectively used for applications such as image magnification and confidential data concealment.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제3권4호
• /
• pp.366-375
• /
• 2009

In collaborative filtering, many neighbors are needed to improve the quality and stability of the recommendation. The quality may not be good mainly due to the high similarity between two users not guaranteeing the same preference for products considered for recommendation. This paper proposes a consistency definition, rather than similarity, based on information entropy between two users to improve the recommendation. This kind of consistency between two users is then employed as a trust metric in collaborative filtering methods that select neighbors based on the metric. Empirical studies show that such collaborative filtering reduces the number of neighbors required to make the recommendation quality stable. Recommendation quality is also significantly improved.

• 대한전자공학회논문지TE
• /
• 제42권4호
• /
• pp.21-26
• /
• 2005

워터마크를 삽입한 동영상의 화질을 향상시키고 다양한 공격에 강인하게 하기 위해 하나의 장면내에서 프레임간의 유사성을 이용하여 워터마크를 삽입하는 방법을 제안한다. 3개의 프레임 단위로 2 프레임씩 짝을 이루어 워터마크를 삽입 검출한다. 실험 결과, 제안한 방법은 PSNR(peak signal to noise ratio) 값이 평균 45dB 정도의 고화질을 나타내며 동영상 압축, 저주파 필터 공격과 프레임 삭제 등의 동영상 편집 공격에서도 강인함을 나타낸다.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2002년도 합동 추계학술대회 논문집 정보 및 제어부문
• /
• pp.206-209
• /
• 2002

This paper proposes a fractal dimension method for measurement of degree of similarity between prototype pattern and input pattern at ART1 (Adaptive Resonance Theory 1) neural network. In order to confirm the validity of proposed method, comparison of the performance has made between the conventional method and the proposed method through simulation. The results show that the proposed method has considerably improved the performance.