• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제7권3호
• /
• pp.522-538
• /
• 2013

Content based image retrieval has become an increasingly important research topic for its wide application. It is highly challenging when facing to large-scale database with large variance. The retrieval systems rely on a key component, the predefined or learned similarity measures over images. We note that, the similarity measures can be potential improved if the data distribution information is exploited using a more sophisticated way. In this paper, we propose a similarity measure learning approach for image retrieval. The similarity measure, so called Fisher kernel, is derived from the probabilistic distribution of images and is the function over observed data, hidden variable and model parameters, where the hidden variables encode high level information which are powerful in discrimination and are failed to be exploited in previous methods. We further propose a discriminative learning method for the similarity measure, i.e., encouraging the learned similarity to take a large value for a pair of images with the same label and to take a small value for a pair of images with distinct labels. The learned similarity measure, fully exploiting the data distribution, is well adapted to dataset and would improve the retrieval system. We evaluate the proposed method on Corel-1000, Corel5k, Caltech101 and MIRFlickr 25,000 databases. The results show the competitive performance of the proposed method.

• 한국지능시스템학회:학술대회논문집
• /
• 한국퍼지및지능시스템학회 1998년도 The Third Asian Fuzzy Systems Symposium
• /
• pp.288-295
• /
• 1998

This paper proposes a new data classification method based on the tolerant rough set that extends the existing equivalent rough set. Similarity measure between two data is described by a distance function of all constituent attributes and they are defined to be tolerant when their similarity measure exceeds a similarity threshold value. The determination of optimal similarity theshold value is very important for the accurate classification. So, we determine it optimally by using the genetic algorithm (GA), where the goal of evolution is to balance two requirements such that (1) some tolerant objects are required to be included in the same class as many as possible. After finding the optimal similarity threshold value, a tolerant set of each object is obtained and the data set is grounded into the lower and upper approximation set depending on the coincidence of their classes. We propose a two-stage classification method that all data are classified by using the lower approxi ation at the first stage and then the non-classified data at the first stage are classified again by using the rough membership functions obtained from the upper approximation set. We apply the proposed classification method to the handwritten numeral character classification. problem and compare its classification performance and learning time with those of the feed forward neural network's back propagation algorithm.

• 한국인터넷방송통신학회논문지
• /
• 제20권4호
• /
• pp.17-22
• /
• 2020

현재 많은 연구가 진행되고 있는 행위 기반 인증 기술은 다른 인증 기술들에 비해서 인증의 인식률을 높이는데 많은 데이터의 장기간 추출이 필요하다. 본 논문은 안드로이드 환경의 스마트폰에 내재되어있는 터치 센서와 자이로스코프를 이용하여 그동안의 행위 기반 인증 연구에서 사용 되었던 행위 특징 데이터들 중에서 핵심적인 최소한의 데이터들만을 이용하기 위해 사용자에게 다섯 차례의 측정을 요구하여 다섯 번의 터치스크린 화면을 터치 하는 방식으로 총 6가지의 행위 특징 데이터를 수집하였고 다음 터치 측정으로 넘어가는 동안의 데이터들의 변화 값에 평균 값을 구하여 이 값과 측정값의 코사인 유사도 측정을 수행하여 코사인 유사도 허용 범위를 생성 한 후, 인증 시도 데이터의 코사인 유사도 값과 비교하는 방식의 사용자 행위 기반 인증 방식을 제안한다. 본 논문을 통해서 적은 수의 특징 데이터와 실험자수 환경에서도 코사인 유사도 인증 범위에 적용되는 임계값을 조절하는 방식을 통해서 최초 EER 37.6%에서 최종 EER 1.9%의 높은 성능을 증명하는데 성공하였다.

• Journal of Pharmaceutical Investigation
• /
• 제33권3호
• /
• pp.245-253
• /
• 2003

Dissolution profile comparsions can be done by virtue of the similarity factor $(f_2)$. It is a logarithmic reciprocal square root transformation of the sum of squared error of % dissolution differences between two profiles at several time points. It gives information on the degree of similarity between the two profiles: An $f_2$ value between 50 and 100 suggests the similarity/equivalence of the two dissolution curves being compared. The objective of this report was to provide a careful examination on the $f_2$ metrics in detail. It was shown that $f_2$ values exceeded 50, when relative differences in % dissolved between two products were less than 15% at all time points. The similarity factor value was also found to be greater than 50, in cases when absolute % dissolution differences were below 10% at all time points. Interestingly, the $f_2$ value was changed by the number of the time points selected for calculation. In particular, $f_2$ tended to have higher values, when the $f_2$ metrics used a large number of time points in which % dissolved reached plateau. Finally, since the similarity factor was a sample statistics, it was impossible to infer type I/II errors and sampling error. Despite certain limitations inherited in the $f_2$ metrics, it was easy and convenient to evaluate how similar the two dissolution profiles were.

• Journal of the Korean Data and Information Science Society
• /
• 제17권4호
• /
• pp.1151-1160
• /
• 2006

The similarity weight, the pearson's correlation coefficient, which is used in the recommender system has a weak point that it cannot predict all of the prediction value. The similarity weight, the vector similarity, has a weak point of the high MAE although the prediction coverage using the vector similarity is higher than that using the pearson's correlation coefficient. The purpose of this study is to suggest how to raise the prediction coverage. Also, the MAE using the suggested method in this study was compared both with the MAE using the pearson's correlation coefficient and with the MAE using the vector similarity, so was the prediction coverage. As a result, it was found that the low of the MAE in the case of using the suggested method was higher than that using the pearson's correlation coefficient. However, it was also shown that it was lower than that using the vector similarity. In terms of the prediction coverage, when the suggested method was compared with two similarity weights as I mentioned above, it was found that its prediction coverage was higher than that pearson's correlation coefficient as well as vector similarity.

• 한국측량학회지
• /
• 제30권2호
• /
• pp.143-152
• /
• 2012

항공영상을 이용하여 수치표면자료와 같은 3차원 자료를 자동으로 제작하기 위해서는 영상정합이 반드시 필요하다. 최근 사용되고 있는 항공 디지털 프레임 영상은 과거의 아날로그 영상에 비해 폐색지역이 적은 고중복도 다중 스트립 영상으로 촬영되기에 용이하다. 최근 다중 스트립 영상을 이용한 다중영상정합 기법에 대한 연구가 많이 이루어지고 있으며, 특히 각 영상에서 추출된 점(point feature)이나 형상(linear feature)의 유사성 측정 방법에 대한 연구가 진행되고 있다. 본 연구에서는 수직궤적 기반 다중영상정합을 대상으로 영역기반 유사성 측정 방법으로 SNCC(Sum of Normalized Cross-Correlation)와 SSD(Sum of Squared-Difference) 방법을 비교 분석하였다. 또한 영역기반 유사성 측정에 필요한 요소로 영상의 화소값, 화소값 기울기 강도, 화소값과 화소값 기울기 강도 평균을 각각 사용하여 결과를 비교하였다. 이 외에도 영역기반 유사성 측정에서 중요한 요소인 기준 윈도우의 크기를 비정규 적응형 기준 윈도우 방법과 정규 적응형 윈도우 방법을 적용하여 결과를 비교 분석하였다. 실험을 위하여 사용된 항공영상은 ZI Imaging 사의 DMC (Digital Modular Camera)에 의해 종중복도는 80%, 횡중복도는 60%로 촬영되었으며, 3개의 스트립으로 구성되었다. 다양한 방법으로 실험을 수행한 결과에 따르면 유사성 측정 방법으로는 SNCC, 유사성 측정 요소로는 화소값과 화소값 기울기 강도 평균, 그리고 비정규 적응형 기준 윈도우가 수직궤적 기반 다중영상정합의 영역기반 유사성 측정에 가장 적합하다는 것을 확인하였다.

• 통상정보연구
• /
• 제11권1호
• /
• pp.139-155
• /
• 2009

This study investigates empirically the export competitiveness of Korea and Japan in America by calculating 4 indexes such as market share index(MSI), export similarity index(ESI), market comparative adventage index(MCAI) and market share expansion ratio(MSER)-export similarity deepening ratio(ESDR). The empirical finding of this analysis shows that Korea is competitive in the labor-intensive products and Japan in the technology-intensive products. This result also meets the general understandings that Japan is superior to Korea in the export competitiveness such as value added of goods, etc. Therefore, in order to strengthen the export competitiveness of Korea in the US market, it's desirable for our firms and government to improve the quality of product ranges by developing technologies focused on the higher value-added products.

• 대한원격탐사학회지
• /
• 제36권1호
• /
• pp.55-65
• /
• 2020

Image classification needs the spectral similarity comparison between spectral features of each pixel and the representative spectral features of each class. The spectral similarity is obtained by computing the spectral feature vector distance between the pixel and the class. Each spectral feature contributes differently in the image classification depending on the class separability of the spectral feature, which is computed using a suitable vector distance measure such as the Bhattacharyya distance. We propose a method to determine the weight value of each spectral feature in the computation of feature vector distance for the similarity measurement. The weight value is determined by the ratio between each feature separability value to the total separability values of all the spectral features. We created ten spectral features consisting of seven bands of Landsat-8 OLI image and three indices, NDVI, NDWI and NDBI. For three experimental test sites, we obtained the overall accuracies between 95.0% and 97.5% and the kappa coefficients between 90.43% and 94.47%.

• 전기학회논문지
• /
• 제65권10호
• /
• pp.1746-1754
• /
• 2016

In this paper, we provide a development and evaluation method for an archery posture analyzing system, using an inertial sensor. The system was developed using LabVIEW2014 by National Instruments and evaluated using the DTW algorithm. To convert the voltage value of the inertial sensor into a physical value, a coordinate transformation matrix bias was applied. To evaluate the similarity of movement in archery shooting, the DTW distance was calculated and similarity was confirmed based on simple mechanical movement, the same person's shooting movement, shooting movement with another person, and the noise signal. The average similarity comparison results were as follows: simple mechanical movement was 17.05%, the same person's shooting movement was 26.48%, shooting movement with another person was 62.8%, and the noise signal was 328.5%; a smaller value indicates a higher level of similarity. We confirmed the possibility of analyzing the archery posture using 3-axis acceleration of the inertial sensor. We inferred that the proposed method might be important means for assessing shooting skills, evaluation of archer's progress, and finding talented archers in advance.

• Journal of Multimedia Information System
• /
• 제5권2호
• /
• pp.67-78
• /
• 2018

This article proposes the modified KNN (K Nearest Neighbor) algorithm which considers the feature similarity and is applied to the word categorization. The texts which are given as features for encoding words into numerical vectors are semantic related entities, rather than independent ones, and the synergy effect between the word categorization and the text categorization is expected by combining both of them with each other. In this research, we define the similarity metric between two vectors, including the feature similarity, modify the KNN algorithm by replacing the exiting similarity metric by the proposed one, and apply it to the word categorization. The proposed KNN is empirically validated as the better approach in categorizing words in news articles and opinions. The significance of this research is to improve the classification performance by utilizing the feature similarities.