• 한국산학기술학회:학술대회논문집
• /
• 한국산학기술학회 2003년도 Proceeding
• /
• pp.223-230
• /
• 2003

Feature-based similarity retrieval has become an important research issue in multimedia database systems. The features of multimedia data are useful for discriminating between multimedia objects (e 'g', documents, images, video, music score, etc.). For example, images are represented by their color histograms, texture vectors, and shape descriptors, and are usually high-dimensional data. The performance of conventional multidimensional data structures(e'g', R- Tree family, K-D-B tree, grid file, TV-tree) tends to deteriorate as the number of dimensions of feature vectors increases. The R＊-tree is the most successful variant of the R-tree. In this paper, we propose a SOM-based R＊-tree as a new indexing method for high-dimensional feature vectors.The SOM-based R＊-tree combines SOM and R＊-tree to achieve search performance more scalable to high dimensionalities. Self-Organizing Maps (SOMs) provide mapping from high-dimensional feature vectors onto a two dimensional space. The mapping preserves the topology of the feature vectors. The map is called a topological of the feature map, and preserves the mutual relationship (similarity) in the feature spaces of input data, clustering mutually similar feature vectors in neighboring nodes. Each node of the topological feature map holds a codebook vector. A best-matching-image-list. (BMIL) holds similar images that are closest to each codebook vector. In a topological feature map, there are empty nodes in which no image is classified. When we build an R＊-tree, we use codebook vectors of topological feature map which eliminates the empty nodes that cause unnecessary disk access and degrade retrieval performance. We experimentally compare the retrieval time cost of a SOM-based R＊-tree with that of an SOM and an R＊-tree using color feature vectors extracted from 40, 000 images. The result show that the SOM-based R＊-tree outperforms both the SOM and R＊-tree due to the reduction of the number of nodes required to build R＊-tree and retrieval time cost.

• 정보처리학회논문지D
• /
• 제8D권5호
• /
• pp.507-512
• /
• 2001

특징 기반 유사성은 멀티미디어 데이터베이스 시스템에서 중요한 연구 쟁점이 되고 있다. 멀티미디어 데이터의 특징이 멀티미디어 객체들을 구별하는데 유용하다지만 특징 벡터의 차원의 수가 증가함에 따라 종래의 다차원 데이터 구조의 성능은 떨어지는 경향이 있다. $R^{*}-Tree$는 R-Tree의 가장 성공적인 병형으로 본 논문에서 고차원 특징 벡터를 위한 새로운 인덱싱 방법으로서 자기 조직화 맵 기반 $R^{*}-Tree$를 제안한다. 자기 조직화 맵 기잔 $R^{*}-Tree$는 고차원 데이터를 좀더 스칼라화해서 탐색할 수 있도록 SOM과 $R^{*}-Tree$를 결합하여 구축한 인덱싱 기법이다. 자기 조직 맵은 고차원 특징 벡터들로부터 2차원 공간으로의 맵핑을 제공한다. 그러나 맵을 위상 특징 맵이라 하고 인접 노느에서 서로 유사한 특징 벡터들을 모아서 입력데이터의 특징 공간들 속에 유사성을 보존하는데 위상 특징 맵의 각 노드는 코드북 벡터를 가지고 있다. 실험적으로 4만개의 이미지로부터 추출된 색깔 특징 벡터들을 이용하여 자기 조직화 맵 기반 $R^{*}-Tree$의 검색시간 비용과 자기 조직화 맵과 $R^{*}-Tree$의 검색 시간 비용을 비교한다. 그 결과 $R^{*}-Tree$를 구축하는데 필요한 노드 수와 검색 시간 비용이 감소됨으로써 자기 조직화 맵 기반 $R^{*}-Tree$는 자기 조직화 맵과 $R^{*}-Tree$보다 훨씬 우수한 성능을 나타냄이 입증되었다.

• 정보처리학회논문지:소프트웨어 및 데이터공학
• /
• 제2권1호
• /
• pp.43-48
• /
• 2013

Adaboost 알고리즘은 얼굴인식을 위한 Haar-like feature들을 이용하기 위해 가장 널리 쓰이고 있는 알고리즘이다. 매우 빠르며 효율적인 성능을 보이고 있으며 하나의 모델이미지가 존재하는 단일분포 데이터에 대해 매우 효율적이다. 그러나 정면 얼굴과 측면 얼굴을 혼합한 인식 등 둘 이상의 모델이미지를 가진 다중 분포모델에 대해서는 그 성능이 저하된다. 이는 단일 학습 알고리즘의 선형결합에 의존하기 때문에 생기는 현상이며 그 응용범위의 한계를 지니게 된다. 본 연구에서는 이를 해결하기 위한 제안으로서 Decision Tree를 Harr-like Feature와 결합하는 기법을 제안한다. Decision Tree를 사용 함으로서 보다 넓은 분야의 문제를 해결하기 위해 기존의 Decision Tree를 Harr-like Feature에 적합하도록 개선한 HDCT라고 하는 Harr-like Feature를 활용한 Decision Tree를 제안하였으며 이것의 성능을 Adaboost와 비교 평가하였다.

• 대한산업공학회지
• /
• 제41권4호
• /
• pp.338-343
• /
• 2015

We suggest a method to configure feature functions of continuous conditional random field (C-CRF). Regression tree and similarity analysis are introduced to construct the first and second feature functions of C-CRF, respectively. Rules from the regression tree are transformed to logic functions. If a logic in the set of rules is true for a data then it returns the corresponding value of leaf node and zero, otherwise. We build an Euclidean similarity matrix to define neighborhood, which constitute the second feature function. Using two feature functions, we make a C-CRF model and an illustrate example is provided.

• 전자공학회논문지B
• /
• 제29B권11호
• /
• pp.102-110
• /
• 1992

In this paper, a new method of character recognition is proposed. It uses density information, in addition to positional and directional information generally used, to recognize a character. Four directional feature primitives are extracted from the thinning templates on the observation that the output of the templates have directional property in general. A simple and fast feature extraction scheme is possible. Features are organized from recursive nonary tree(N-tree) that corresponds to normalized character area. Each node of the N-tree has four directional features that are sum of the features of it's nine sub-nodes. Every feature primitive from the templates are added to the corresponding leaf and then summed to the upper nodes successively. Recognition can be accomplished by using appropriate feature level of N-tree. Also, effectiveness of each node's feature vector was tested by experiment. A method to implement the proposed feature vector organization algorithm into hardware is proposed as well. The third generation node, which is 4$\times$4, is used as a unit processing element to extract features, and it was implemented in hardware. As a result, we could observe that it is possible to extract feature vector for real-time processing.

• Journal of Power Electronics
• /
• 제8권1호
• /
• pp.101-107
• /
• 2008

In this paper, the features extracted from vibration time signals are used to detect the bearing fault condition. The decision tree is applied to diagnose the bearing status, which has the benefits of being an expert system that is based on knowledge history and is simple to understand. This paper also suggests a genetic algorithm (GA) as a method to reduce the number of features. In order to show the potentials of this method in both aspects of accuracy and simplicity, the reduced-feature decision tree is compared with the non reduced-feature decision tree and the PCA-based decision tree.

• Communications for Statistical Applications and Methods
• /
• 제2권1호
• /
• pp.74-88
• /
• 1995

This article introduces a unified method of choosing the most explanatory and significant multiway partitions for classification tree design and analysis. The method is derived on the impurity reduction (IR) measure of divergence, which is proposed to extend the proportional-reduction-in-error (PRE) measure in the decision-theory context. For the method derivation, the IR measure is analyzed to characterize its statistical properties which are used to consistently handle the subjects of feature formation, feature selection, and feature deletion required in the associated classification tree construction. A numerical example is considered to illustrate the proposed approach.

• 한국지능시스템학회:학술대회논문집
• /
• 한국퍼지및지능시스템학회 1995년도 추계학술대회 학술발표 논문집
• /
• pp.218-223
• /
• 1995

A new design scheme of a binary decision tree is proposed. In this scheme a binary decision tree is constructed by using genetic algorithm and FCM algorithm. At each node optimal or near-optimal feature or feature subset among all the available features is selected based on fitness function in genetic algorithm which is inversely proportional to classification error, balance between cluster, number of feature used. The proposed design scheme is applied to the handwtitten alphabetic characters. Experimental results show the usefulness of the proposed scheme.

• 한국HCI학회:학술대회논문집
• /
• 한국HCI학회 2007년도 학술대회 1부
• /
• pp.667-674
• /
• 2007

Facial expressions provide significant clues about one's emotional state; however, it always has been a great challenge for machine to recognize facial expressions effectively and reliably. In this paper, we report a method of feature-based adaptive motion energy analysis for recognizing facial expression. Our method optimizes the information gain heuristics of ID3 tree and introduces new approaches on (1) facial feature representation, (2) facial feature extraction, and (3) facial feature classification. We use minimal reasonable facial features, suggested by the information gain heuristics of ID3 tree, to represent the geometric face model. For the feature extraction, our method proceeds as follows. Features are first detected and then carefully "selected." Feature "selection" is finding the features with high variability for differentiating features with high variability from the ones with low variability, to effectively estimate the feature's motion pattern. For each facial feature, motion analysis is performed adaptively. That is, each facial feature's motion pattern (from the neutral face to the expressed face) is estimated based on its variability. After the feature extraction is done, the facial expression is classified using the ID3 tree (which is built from the 1728 possible facial expressions) and the test images from the JAFFE database. The proposed method excels and overcomes the problems aroused by previous methods. First of all, it is simple but effective. Our method effectively and reliably estimates the expressive facial features by differentiating features with high variability from the ones with low variability. Second, it is fast by avoiding complicated or time-consuming computations. Rather, it exploits few selected expressive features' motion energy values (acquired from intensity-based threshold). Lastly, our method gives reliable recognition rates with overall recognition rate of 77%. The effectiveness of the proposed method will be demonstrated from the experimental results.

• 정보처리학회논문지D
• /
• 제8D권4호
• /
• pp.305-312
• /
• 2001

최근 고차원 색인 구조들이 멀티미디어 데이터베이스, 데이터 웨어하우징과 같은 데이터베이스 응용에서 유사성 검색을 위해 요구된다. 본 논문에서는 고차원 특징벡터에 대한 효율적인 저장과 검색을 지원하는 셀-기반 시그니쳐 트리(CS-트리)를 제안한다. 제안하는 CS-트리는 고차원 특징 벡터 공간을 셀로써 분할하여 하나의 특징 벡터를 그에 해당되는 셀의 시그니쳐로 표현한다. 특징 벡터 대신 셀의 시그니쳐를 사용함으로써 트리의 깊이를 줄이고, 그 결과 효율적인 검색 성능을 달성한다. 또한 셀에 기반하여 탐색 공간을 효율적으로 줄이는 유사성 검색 알고리즘을 제시한다. 마지막으로 우수한 고차원 색인 기법으로 알려져 있는 X-트리와 삽입시간, k-최근접 질의에 대한 검색 시간 그리고 부가저장 공간 측면에서 성능 비교를 수행한다. 성능비교 결과 CS-트리가 검색 성능에서 우수함을 보인다.