• The Journal of the Korea institute of electronic communication sciences
• /
• v.11 no.7
• /
• pp.677-684
• /
• 2016

With the increase of multimedia information such as image, researches on extracting high-level semantic information from low-level visual information has been realized, and in order to automatically generate this kind of information. Various technologies have been developed. Generally, image retrieval is widely preceded by comparing colors and shapes among images. In some cases, images with similar color, shape and even meaning are hard to retrieve. In this article, in order to retrieve the object in an image, technical value of middle level is converted into meaning value of middle level. Furthermore, to enhance accuracy of segmentation, K-means algorithm is engaged to compute k values for various images. Thus, object retrieval can be achieved by segmented low-level feature and relationship of meaning is derived from ontology. The method mentioned in this paper is supposed to be an effective approach to retrieve images as required by users.

• KIPS Transactions on Software and Data Engineering
• /
• v.3 no.7
• /
• pp.285-292
• /
• 2014

A named entity recognition method is used to improve the performance of information retrieval systems, question answering systems, machine translation systems and so on. The targets of the named entity recognition are usually PLOs (persons, locations and organizations). They are usually proper nouns or unregistered words, and traditional named entity recognizers use these characteristics to find out named entity candidates. The titles of books, movies and TV programs have different characteristics than PLO entities. They are sometimes multiple phrases, one sentence, or special characters. This makes it difficult to find the named entity candidates. In this paper we propose a method to quickly extract title named entities from news articles and automatically build a named entity dictionary for the titles. For the candidates identification, the word phrases enclosed with special symbols in a sentence are firstly extracted, and then verified by the SVM with using feature words and their distances. For the classification of the extracted title candidates, SVM is used with the mutual information of word contexts.

• Proceedings of the Korean Society of Surveying, Geodesy, Photogrammetry, and Cartography Conference
• /
• 2010.04a
• /
• pp.301-306
• /
• 2010

Recently GIS is not only displaying and servicing data on the 2D, but also is changing rapidly to display and service 3D data. Also 3D related technology is developing actively. For display of 3D data, terrain DTM has become a basis. Generally, to acquire DTM, users are using LIDAR data or digital map's contour line. However, if using these data for producing DTM, users need to additional cost and data lead time. And hard to update terrain data. For possibility of solving these problem, this study did DTM extraction with automatic matching for aerial photograph, and analysed the result with measurement of Orthometric height and excuted accuracy through DTM(which extracted from digital photogrammetric technique). As a result, we can get a high accuracy of RMSE (0.215m).

• Journal of KIISE:Computing Practices and Letters
• /
• v.15 no.9
• /
• pp.695-699
• /
• 2009

In this paper, we propose an automatic prostate segmentation technique using image intensity and gradient information. Our method is composed of four steps. First, rays at regular intervals are generated. To minimize the effect of noise, the start and end positions of the ray are calculated. Second, the profiles on each ray are sorted based on the gradient. And priorities are applied to the sorted gradient in the profile. Third, boundary points are extracted by using gradient priority and intensity distribution. Finally, to reduce the error, the extracted boundary points are corrected by using B-spline interpolation. For accuracy evaluation, the average distance differences and overlapping region ratio between results of manual and automatic segmentations are calculated. As the experimental results, the average distance difference error and standard deviation were 1.09mm $\pm0.20mm$. And the overlapping region ratio was 92%.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.40 no.4
• /
• pp.666-677
• /
• 2015

The importance of application traffic analysis for efficient network management has been emphasized continuously. Snort is a popular traffic analysis system which detects traffic matched to pre-defined signatures and perform various actions based on the rules. However, it is very difficult to get highly accurate signatures to meet various analysis purpose because it is very tedious and time-consuming work to search the entire traffic data manually or semi-automatically. In this paper, we propose a novel method to generate signatures in a fully automatic manner in the form of sort rule from raw packet data captured from network link or end-host. We use a sequence pattern algorithm to generate common substring satisfying the minimum support from traffic flow data. Also, we extract the location and header information of the signature which are the components of snort content rule. When we analyzed the proposed method to several application traffic data, the generated rule could detect more than 97 percentage of the traffic data.

• The KIPS Transactions:PartB
• /
• v.14B no.4
• /
• pp.295-302
• /
• 2007

It is very iterative and complicated work to enter the personal information every time one fills the form-based resume or one joins the new membership page on the internet. Although there are some systems that have the personal information on the computer and fill the membership page automatically, their accuracies are not often satisfactory in that the fields and their values do not match exactly. The research proposes and implements a system that has user's information on the computer and reasons and fills the information automatically that a membership web page(target page) requests using the personal information ontology. During the reasoning process, the target page is analyzed to extract the requested fields. Then the requested field names are converted to the standard field names using synonym ontology. The converted requested fields find the appropriate level in the personal information ontology using ontology match making to generate the final field value. The system not only finds the similar fields but also generates the exact field values by reasoning on the information ontology hierarchy. By experimenting with several membership pages on the web, the system showed higher accuracy over the existing systems. The system can be easily applicable to the cases where one iteratively fills the same information such as resume form.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.37 no.6
• /
• pp.545-554
• /
• 2019

This research aims to provide a building area extraction approach over the areas where data acquisition is impossible through field surveying, aerial photography and lidar scanning. Hence, high-resolution satellite images, which have high accessibility over the earth, are utilized for the automated building extraction in this study. 3D point clouds or DSM (Digital Surface Models), derived from the stereo image matching process, provides low quality of building area extraction due to their high level of noises and holes. In this regards, this research proposes a hybrid building area extraction approach which utilizes 3D point clouds (from image matching), and color and linear information (from imagery). First of all, ground and non-ground points are separated from 3D point clouds; then, the initial building hypothesis is extracted from the non-ground points. Secondly, color based building hypothesis is produced by considering the overlapping between the initial building hypothesis and the color segmentation result. Afterwards, line detection and space partitioning results are utilized to acquire the final building areas. The proposed approach shows 98.44% of correctness, 95.05% of completeness, and 1.05m of positional accuracy. Moreover, we see the possibility that the irregular shapes of building areas can be extracted through the proposed approach.

• Journal of KIISE:Software and Applications
• /
• v.33 no.2
• /
• pp.143-153
• /
• 2006

Microarray which enables us to obtain hundreds and thousands of expression of gene or genotype at once is an epoch-making technology in comparative analysis of genes. First of all, we have to measure the intensity of each gene in an microarray image from the experiment to gain the expression level of each gene. But it is difficult to analyze the microarray image in manual because it has a lot of genes. Meta-gridding method and various auto-gridding methods have been proposed for this, but thew still have some problems. For example, meta-gridding requires manual-work due to some variations in spite of experiment in same microarray, and auto-gridding nay not carried out fully or correctly when an image has a lot of noises or is lowly expressed. In this article, we propose Hierarchical Grid Alignment algorithm for new methodology combining meta-gridding method with auto-gridding method. In our methodology, we necd a meta-grid as an input, and then align it with the microarray image automatically. Experimental results show that the proposed method serves more robust and reliable gridding result than the previous methods. It is also possible for user to do more reliable batch analysis by using our algorithm.

• Journal of Internet Computing and Services
• /
• v.7 no.2
• /
• pp.23-35
• /
• 2006

This work presents o novel method which extract facial region und features from motion picture automatically and controls the 3D facial expression in real time. To txtract facial region and facial feature points from each color frame of motion pictures a new nonparametric skin color model is proposed rather than using parametric skin color model. Conventionally used parametric skin color models, which presents facial distribution as gaussian-type, have lack of robustness for varying lighting conditions. Thus it needs additional work to extract exact facial region from face images. To resolve the limitation of current skin color model, we exploit the Hue-Tint chrominance components and represent the skin chrominance distribution as a linear function, which can reduce error for detecting facial region. Moreover, the minimal facial feature positions detected by the proposed skin model are adjusted by using edge information of the detected facial region along with the proportions of the face. To produce the realistic facial expression, we adopt Water's linear muscle model and apply the extended version of Water's muscles to variation of the facial features of the 3D face. The experiments show that the proposed approach efficiently detects facial feature points and naturally controls the facial expression of the 3D face model.

• Smart Media Journal
• /
• v.8 no.2
• /
• pp.58-65
• /
• 2019

Automatic text summarization is a process of shortening a text document by either extraction or abstraction. The abstraction approach inspired by deep learning methods scaling to a large amount of document is applied in recent work. Abstractive text summarization involves utilizing pre-generated word embedding information. Low-frequent but salient words such as terminologies are seldom included to dictionaries, that are so called, out-of-vocabulary(OOV) problems. OOV deteriorates the performance of Encoder-Decoder model in neural network. In order to address OOV words in abstractive text summarization, we propose a copy mechanism to facilitate copying new words in the target document and generating summary sentences. Different from the previous studies, the proposed approach combines accurate pointing information and selective copy mechanism based on bidirectional RNN and bidirectional LSTM. In addition, neural network gate model to estimate the generation probability and the loss function to optimize the entire abstraction model has been applied. The dataset has been constructed from the collection of abstractions and titles of journal articles. Experimental results demonstrate that both ROUGE-1 (based on word recall) and ROUGE-L (employed longest common subsequence) of the proposed Encoding-Decoding model have been improved to 47.01 and 29.55, respectively.