• Journal of KIISE:Databases
• /
• v.35 no.5
• /
• pp.379-390
• /
• 2008

Due to the increase of the amount of image data, the demand for searching similar images is continuously increasing. Therefore, many researches about the content-based image retrieval (CBIR) are conducted to search similar images effectively. In CBIR, it uses image contents such as color, shape, and texture for more effective retrieval. However, when we apply CBIR to satellite images which are complex and pose the difficulty in using color information, we can have trouble to get a good retrieval result. Since it is difficult to use color information of satellite images, we need image segmentation to use shape information by separating the shape of an object in a satellite image. However, because satellite images are complex, image segmentation is hard and poor image segmentation results in poor retrieval results. In this paper, we propose a new approach to search similar images without image segmentation for satellite images. To do a similarity search without image segmentation, we define a similarity of an image by considering SIFT keypoint descriptors which doesn't require image segmentation. Experimental results show that the proposed approach more effectively searches similar satellite images which are complex and pose the difficulty in using color information.

• Journal of applied mathematics & informatics
• /
• v.9 no.2
• /
• pp.793-806
• /
• 2002

Feature-based similarity retrieval become an important research issue in image database systems. The features of image data are useful to discrimination of images. In this paper, we propose the highspeed k-Nearest Neighbor search algorithm based on Self-Organizing Maps. Self-Organizing Maps (SOM) provides a mapping from high dimensional feature vectors onto a two-dimensional space. The mapping preserves the topology of the feature vectors. The map is called topological feature map. A topological feature map preserves the mutual relations (similarity) in feature spaces of input data. and clusters mutually similar feature vectors in a neighboring nodes. Each node of the topological feature map holds a node vector and similar images that is closest to each node vector. In topological feature map, there are empty nodes in which no image is classified. We experiment on the performance of our algorithm using color feature vectors extracted from images. Promising results have been obtained in experiments.

• Journal of KIISE:Software and Applications
• /
• v.37 no.2
• /
• pp.137-147
• /
• 2010

A blog is a website, usually maintained by an individual, with regular entries of commentary, descriptions of events, or other material such as graphics or video. Entries are commonly displayed in reverse chronological order. Blog search engines, like web search engines, seek information for searchers on blogs. Blog search engines sometimes output unsatisfactory results, mainly due to spam blogs or splogs. Splogs are blogs hosting spam posts, plagiarized or auto-generated contents for the sole purpose of hosting advertizements or raising the search rankings of target sites. This thesis focuses on splog detection. This thesis proposes a new splog detection method, which is based on blog post structure similarity and posting count per day. Experiments based on methods proposed a day show excellent result on splog detection tasks with over 90% accuracy.

• Journal of Korean Library and Information Science Society
• /
• v.40 no.4
• /
• pp.139-158
• /
• 2009

This paper used the network analysis method to analyse a variety of attributes of searcher's search behaviors which was appeared on search access log data. The results of this research are as follows. First, the structure of network represented depending on the similarity of the query that user had inputed. Second, we can find out the particular searchers who occupied in the central position in the network. Third, it showed that some query were shared with ego-searcher and alter searchers. Fourth, the total number of searchers can be divided into some sub-groups through the clustering analysis. The study reveals a new recommendation algorithm of associated searchers and search query through the social network analysis, and it will be capable of utilization.

• Journal of KIISE:Databases
• /
• v.27 no.2
• /
• pp.199-208
• /
• 2000

The k-nearest neighbor search query based on similarity is very important for content-based multimedia information retrieval(MIR). The conventional exact k-nearest neighbor search algorithm is not efficient for the MIR application because multimedia data should be represented as high dimensional feature vectors. Thus, an approximate k-nearest neighbor search algorithm is required for the MIR applications because the performance increase may outweigh the drawback of receiving approximate results. For this, we propose a new approximate k-nearest neighbor search algorithm for high dimensional data. In addition, the comparison of the conventional algorithm with our approximate k-nearest neighbor search algorithm is performed in terms of retrieval performance. Results show that our algorithm is more efficient than the conventional ones.

• The Journal of the Korea Contents Association
• /
• v.11 no.5
• /
• pp.125-137
• /
• 2011

Recently the research on the system using images taken from camera phones as input is actively conducted. This paper proposed a system that shows wine pictures which are similar to the input wine label in order. For the calculation of the similarity of images, the representative color of each cell of the image, the recognized text color, background color and distribution of feature points are used as the features. In order to calculate the difference of the colors, RGB is converted into CIE-Lab and the feature points are extracted by using Harris Corner Detection Algorithm. The weights of representative color of each cell of image, text color and background color are applied. The image similarity is calculated by normalizing the difference of color similarity and distribution of feature points. After calculating the similarity between the input image and the images in the database, the images in Database are shown in the descent order of the similarity so that the effort of users to search for similar wine labels again from the searched result is reduced.

• Korean Journal of Computational Design and Engineering
• /
• v.11 no.4
• /
• pp.315-325
• /
• 2006

It is very often necessary to search for similar parts during designing a new product because its parts are often easily designed by modifying existing similar parts. In this way, the design time and cost can be reduced. Thus it would be nice to have an efficient similarity comparison algorithm that can be used anytime in the design process. There have been many approaches to compare shape similarity between two solids. In this paper, two parts represented in B-Rep is compared in two steps: one for overall appearances and the other for detail features. In the first step, geometric information is used in low level of detail for easy and fast pre-classification by the overall appearance. In the second step, feature information is used to compare the detail shape in high level of detail to find more similar design. To realize the idea above, a multi resolution algorithm is proposed so that a given solid is described by an overall appearance in a low resolution and by detail features in high resolution. Using this multi-resolution representation, parts can be compared based on the overall appearance first so that the number of parts to be compared in high resolution is reduced, and then detail features are investigated to retrieve the most similar part. In this way, computational time can be reduced by the fast classification in the first step while reliability can be preserved by detail comparison in the second step.

• Journal of the Korea Society of Computer and Information
• /
• v.24 no.6
• /
• pp.57-65
• /
• 2019

In this paper, we study on finding the rail space in elevators by analyzing each image captured with CCD camera. We propose a method that applies one-dimensional matched filter to the pixels of a selected search space in the vertical line at a horizontal position and decides the position with the thickness of the space being represented by a black thick line in captured images. The pattern similarity representing how strongly the associated image pixels resemble with the thick line is defined and calculated with respect to each position along the vertical line of pixels. The position and thickness of the line are decided from the point having the maximum in pattern similarity graph. In the experiments of the proposed method under different illuminational conditions, it is observed that all the pattern similarity graphs show similar shape around door area independent of the conditions and the method can effectively detect the rail space if the rails are illuminated with even weak light. The method can be used for real-time embedded systems because of its simple algorithm, in which it is implemented in simple structure of program with small amount of operations in comparison with the conventional approaches using Canny edge detection and Hough transform.

• Journal of the Korea Society of Computer and Information
• /
• v.29 no.5
• /
• pp.1-10
• /
• 2024

Attackers tend to use similar vulnerabilities when finding their next target IT assets. They also continuously search for new attack targets. Therefore, it is essential to find the potential targets of attackers in advance. Our method proposes a novel approach for efficient vulnerable asset management and zero-day response. In this paper, we propose the ability to detect the IT assets that are potentially infected by the recently discovered vulnerability based on clustering and similarity results. As the experiment results, 86% of all collected assets are clustered within the same clustering. In addition, as a result of conducting a similarity calculation experiment by randomly selecting vulnerable assets, assets using the same OS and service were listed.

• Journal of KIISE:Software and Applications
• /
• v.34 no.1
• /
• pp.1-10
• /
• 2007

This paper proposes a new indexing mechanism for MPEG-7 visual descriptors, especially Dominant Color and Contour Shape descriptors, that guarantees an efficient similarity search for the multimedia database whose visual meta-data are represented with MPEG-7. Since the similarity metric used in the Dominant Color descriptor is based on Gaussian mixture model, the descriptor itself could be transform into a color histogram in which the distribution of the color values follows the Gauss distribution. Then, the transformed Dominant Color descriptor (i.e., the color histogram) is indexed in the proposed indexing mechanism. For the indexing of Contour Shape descriptor, we have used a two-pass algorithm. That is, in the first pass, since the similarity of two shapes could be roughly measured with the global parameters such as eccentricity and circularity used in Contour shape descriptor, the dissimilar image objects could be excluded with these global parameters first. Then, the similarities between the query and remaining image objects are measured with the peak parameters of Contour Shape descriptor. This two-pass approach helps to reduce the computational resources to measure the similarity of image objects using Contour Shape descriptor. This paper also proposes two integration schemes of visual descriptors for an efficient retrieval of multimedia database. The one is to use the weight of descriptor as a yardstick to determine the number of selected similar image objects with respect to that descriptor, and the other is to use the weight as the degree of importance of the descriptor in the global similarity measurement. Experimental results show that the proposed indexing and integration schemes produce a remarkable speed-up comparing to the exact similarity search, although there are some losses in the accuracy because of the approximated computation in indexing. The proposed schemes could be used to build a multimedia database represented in MPEG-7 that guarantees an efficient retrieval.