Dongun An

Detection of human faces using skin color and eyes

We propose an effective and robust automatic face detection method which can locate the faces in natural scene images. The method can be used as a preprocessor of a face recognition system. Several approaches that use the human skin color in the images have been proposed for detecting human faces. They cannot adequately cope with the problem such as false positives, the human skin color of other body parts and face orientation. We use two natural and powerful factors, the skin color and the eyes. This method takes two steps in order to extract regions of the human skin color and the eyes in the images. In the first step of the processing, regions of the human skin color are extracted and those regions are converted into gray scale from the original image. The skin color is a fast and robust cue that can be used as the first step in a face detection. In the next step, the eyes are extracted from candidate regions by calculating central moments and by using geometrical relationships between the eyes. The position and angles between the eyes can determine the exact location of a face. Experimental results using the proposed method show that human faces can be detected in color images regardless of size, expression, lighting conditions and orientation.

Harmful Contents Classification Using the Harmful Word Filtering and SVM

As World Wide Web is more popularized nowadays, it is also creating many problems due to uncontrolled flood of information. The pornographic, violent and other harmful information freely available to the youth, who must be protected by the society, or other users who lack the power of judgment or self-control is creating serious social problems. To resolve those harmful words, various methods proposed and studied. This paper proposes and implements the protecting system that protects internet youth user from harmful contents. To effectively classify harmful/harmless contents, this system uses two steps of classification: harmful word filtering and SVM learning based filtering. We achieved result that the average precision of 92.1%.

Selection of Cluster Hierarchy Depth in Hierarchical Clustering Using K-Means Algorithm

Many papers have shown that the hierarchical clustering method takes good-performance, but is limited because of its quadratic time complexity. In contrast, with a large number of variables, K-means has a time complexity that is linear in the number of documents, but is thought to produce inferior clusters. Think of the factor of simplify, high-quality and high-efficiency, we combine the two approaches providing a new system named CONDOR system with hierarchical structure based on document clustering using K-means algorithm. Evaluated the performance on different hierarchy depth and initial uncertain centroid number based on variational relative document amount correspond to given queries. Comparing with regular method that the initial centroids have been established in advance, our method performance has been improved a lot.

A Study on Antecedent Decision Rules for Natural Language Requirements Document in Korean

In the characteristically requirements document (RD), the pronoun is scarcely used in RD in general. But if we would like to get the higher accuracy in analysis of RD automatically, antecedent decision of pronoun is very important for elicitation of formal requirements (i.e. component, action, statement and parameters etc.) automatically from natural language requirements document (NLRD) via natural language processing (NLP). In this paper, we propose the antecedent decision rule (ADR) to decide antecedent of pronoun from NLRD in Korean via NLP. Pronoun can be classified in personal pronoun and demonstrative pronoun. In the RD, a personal pronoun is almost not occurred so we are focused on antecedent decision for a demonstrative pronoun. The purpose of this paper is to describe the ADR that is used in postprocessing to reduce a burden of morphological analysis and parsing, so our approach can describe additional rules because every rule are applied in the postprocessing. We carried out an experiment with five brief NLRD added pronoun in.

Document Clustering Method using PCA and Fuzzy Association

ABSTRACT This paper proposes a new document clustering method using PCA and fuzzy association. The proposed method can represent an inherent structure of document clusters better since it select the cluster label and terms of representing cluster by semantic features based on PCA. Also it can improve the quality of document clustering because the clustered documents by using fuzzy association values distinguish well dissimilar documents in clusters. The experimental results demonstrate that the proposed method achieves better performance than other document clustering methods.Keywords:Document Clustering, Principal Componet Analysis, Semantic Features, Fuzzy Association 1. 서 론 1) 근래의 정보 검색 분야에는 사용자의 요구사항을 만족시키기 위하여 다양한 정보를 효율적으로 처리할 수 있는 문서의 범주화에 대해서 많은 연구가 있다. 문서의 범주화는 대량의 문서들을 각각의 문서의 특성 및 주제에 맞게 분류하는 것으로, 사전에 학습이 필요한 지도학습방법인 문서분류와 학습이 필요 없는 비지도학습 방법의 문서군집으로 구분할 수 있다[4]. 전통적인 군집방법은 분할기반 방법, 계층적 기반 방법, 밀도기반 방법, 격자 기반 방법으로 분류 할 수 있다. 이들 대부분의 방법들은 거리 기반의 목적 함수를 사용하기 때문에 고차원의 객체들을 군집하는 것에는 비효율적이다. 이중에서 대표적인 군집방법으로는, 군집을 생성하는 방법에 따라서 k개의 군집을 임의로 정하여 군집을 확장해가는 비계

Selection of Cluster Hierarchy Depth and Initial Centroids in Hierarchical Clustering using K-Means Algorithm

Fast and high-quality document clustering algorithms play an important role in providing data exploration by organizing large amounts of information into a small number of meaningful clusters. Many papers have shown that the hierarchical clustering method takes good-performance, but is limited because of its quadratic time complexity. In contrast, with a large number of variables, K-means has a time complexity that is linear in the number of documents, but is thought to produce inferior clusters. In this paper, Condor system using K-Means algorithm Compares with regular method that the initial centroids have been established in advance, our method performance has been improved a lot.