• The KIPS Transactions:PartA
• /
• v.9A no.3
• /
• pp.311-316
• /
• 2002

LZW is a popular variant of LZ78 to compress text documents. LZW yields a high compression rate and is widely used by many commercial programs. Its core idea is to assign most probably used character group an entry in a dictionary. If a group of character which is already positioned in a dictionary appears in the streaming data, then an index of a dictionary is replaced in the position of character group. In this paper, we propose a new efficient method to find least used entries in a dictionary using counter. We also achieve higher compression rate by preassigning widely used tags in hyper-text documents. Experimental results show that the proposed method is more effective than V.42bis and Unix compression method. It gives 3∼8％ better in the standard Calgary Corpus and 23∼24％ better in HTML documents.

• Journal of KIISE:Computer Systems and Theory
• /
• v.32 no.5
• /
• pp.255-259
• /
• 2005

To search a pattern P in a text, such index data structures as suffix trees and suffix arrays are widely used in diverse applications of string processing and computational biology. It is well known that searching in suffix trees is faster than suffix ways in the aspect of time complexity, i.e., it takes O(${\mid}P{\mid}$) time to search P on a constant-size alphabet in a suffix tree while it takes O(${\mid}P{\mid}+logn$) time in a suffix way where n is the length of the text. In this paper we present a linear-tim8 search algorithm in suffix arrays for constant-size alphabets. For a gene.al alphabet $\Sigma$, it takes O(${\mid}P{\mid}log{\mid}{\Sigma}{\mid}$) time.

• 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.

• Journal of KIISE:Databases
• /
• v.34 no.2
• /
• pp.141-152
• /
• 2007

Selectivity estimation has been a crucial component in query optimization in relational databases. While extensive researches have been done on this topic for the predicates of numerical data, only little work has been done for substring predicates. We propose novel suffix tree transform algorithms for this problem. Unlike previous approaches where a full suffix tree is pruned and then an estimation algorithm is employed, we transform a suffix tree into a suffix graph systematically. In our approach, nodes with similar counts are merged while structural information in the original suffix tree is preserved in a controlled manner. We present both an error-bound algorithm and a space-bound algorithm. Experimental results with real life data sets show that our algorithms have lower average relative error than that of the previous works as well as good error distribution characteristics.

• The Journal of the Korea Contents Association
• /
• v.12 no.2
• /
• pp.105-115
• /
• 2012

In this paper, we propose a correction method using phoneme unit segmentation to solve misrecognition of Korean Texts in signboard images using weighted Disassemble Levenshtein Distance. The proposed method calculates distances of recognized texts which are segmented into phoneme units and detects the best matched texts from signboard text database. For verifying the efficiency of the proposed method, a database dictionary is built using 1.3 million words of nationwide signboard through removing duplicated words. We compared the proposed method to Levenshtein Distance and Disassemble Levenshtein Distance which are common representative text string comparison algorithms. As a result, the proposed method based on weighted Disassemble Levenshtein Distance represents an improvement in recognition rates 29.85% and 6% on average compared to that of conventional methods, respectively.

• Proceedings of the Korean Information Science Society Conference
• /
• 2001.04b
• /
• pp.205-207
• /
• 2001

웹 상의 HTML 문서들은 수시로 변경되고 있으며, 정보를 검색하는 웹사이트 또한 예외는 아니다. 다수의 웹 검색엔진들의 결과를 통합하는 메타 검색엔진은 각 검색엔진의 정보 변경에 민감해야 된다. 본 논문은, 수시로 변경되는 검색엔진들의 HTML 문서 정보를 메타 검색 엔진에 반영하기 위해, 자동적으로 검색엔진들의 질의 형태 변경과 검색 엔진의 검색 결과 HTML 문서의 구조 변경 탐지는 질의 결과가 반복되는 HTML 태그(tags) 문서 구조를 패턴(pattern)으로 이용한다. 패턴 발견 알고리즘은 문자열에서 규칙적으로 발생하는 패턴을 찾아내는 Jaak Vilo 알고리즘을 기반으로 HTML 문서를 처리할 수 있도록 확장하였다. 발견된 HTML 문서 패턴과 기존의 검색 엔진 HTML 페이지의 구조적 패턴 정보를 비교하여 문서 구조 변경을 탐지한다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2005.11b
• /
• pp.718-720
• /
• 2005

웹 문서를 대상으로 하는 정보 추출이나 웹 마이닝에 관한 연구가 활발히 진행되면서 특히, 웹에서 나타나는 구조적 패턴을 이용해 정보를 추출하는 방법에 대한 연구가 이루어 지고 있다. 기존의 연구는 HTML을 단순 문자열로 취급하였으나 연구가 거듭됨에 따라 트리로 접근하는 방안에 대해 논의가 되었으며 성능 또한 우수한 것으로 평가되고 있다. 하지만, 기존의 트리 편집 거리의 기법은 모든 노드가 동일한 값을 가진다는 가정하에 진행되는 것으로 HTML의 특성과는 맞지 않다. HTMI은 브라우저에 정보를 보여주기 위한 도구이며 실제 브라우저에 보여지는 내용의 비율이 트리에서의 노드의 비율과 항상 같은 것은 아니기 때문이다. 이 논문에서는 위와 같은 HTML의 특성을 이용하여 노드가 가진 정보의 크기에 따라 서로 다른 비율의 기여도를 부여하고, 이를 고려한 개선된 트리 편집 거리 측정 알고리즘을 이용하여 좀더 나은 패턴 추출 방법을 제안하고자 한다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2003.04c
• /
• pp.395-397
• /
• 2003

온라인 필기 인식기의 필기 모델을 응용하여 오프라인 한글 필기의 필기 궤적을 추적하고 인식하는 방법을 제안한다. 사용한 온라인 모델은 HMM의 망으로 구성한 조합형 한글 필기 모델 BongNet이다. 그리고 시계열 신호의 길이에 대한 모델이 전혀 없는 표준 HMM 대신 동적인 연속 출력 nonstationary HMM 을 이용한 방법을 기술하였다. 획 추적 계산 과정에는 프레임 동기 알고리즘을 적용한다 HMM의 각 상태는 가능한 필기 궤적상의 위치에 대한 정보를 기록한다. 매 시각마다 최종 상태의 후보 중에서 모든 획을 완전히 지나는 경로가 있는지를 조사한다. 본 방법은 문자영상에서 온라인 시계열 코드를 만들어 가는 과정이며 코드와 동시에 인식결과를 출력한다.

• Journal of KIISE:Software and Applications
• /
• v.28 no.11
• /
• pp.813-821
• /
• 2001

While recognition of isolated units of writing, such as a character or a word, has been extensively studied, emphasis on the segmentation itself has been lacking. In this paper we propose an active segmentation method for handwritten Hangul address strings based on the Run-length code. A slant correction algorithm, which is considered as an important preprocessing step for the segmentation, is presented. Three fundamental candidate estimation functions are introduced to detect the clues on touching points, and the classification of touching types is attempted depending on the structural peculiarity of Hangul. Our experiments show segmentation performance of 88.2% on touching characters with minimal over-segmentation.

• Journal of The Korean Association of Information Education
• /
• v.1 no.1
• /
• pp.28-37
• /
• 1997

This paper presents a new neural network based on fuzzy set and its application to invariant character recognition. The fuzzy neural network consists of five layers. The results of simulation show that the network can recognize characters in the case of distortion, translation, rotation and different sizes of handwritten characters and even with noise(8${\sim}$30%)). Translation, distortion, different sizes and noise are achieved by layer L2 and rotation invariant by layer L5. The network can recognize 108 examples of training with 100% recognition rate when they are shifted in eight directions by 1 pixel and 2 pixels. Also, the network can recognize all the distorted characters with 100% recognition rate. The simulations show that the test patterns cover a ${\pm}20^{\circ}$ range of rotation correctly. The proposed network can also recall correctly all the learned characters with 100% recognition rate. The proposed network is simple and its learning and recall speeds are very fast. This network also works for the segmentation and recognition of handwritten characters.