Interactive Data Mining for Large-Scale Image Databases Based on Formal Concept Analysis

被引：5

作者：

Tanabata, Takanari ^{[1
]}

Sawase, Kazuhito ^{[2
]}

Nobuhara, Hajime ^{[2
]}

Bede, Barnabas ^{[3
]}

机构：

[1] Natl Inst Agrobiol Sci, Photobiol & Photosynth Res Unit, 2-1-2 Kannondai, Tsukuba, Ibaraki 3058602, Japan

[2] Univ Tsukuba, Dept Intelligent Interact Technol, Tsukuba, Ibaraki 3058573, Japan

[3] Univ Texas Pan Amer, Dept Math, Edinburg, TX 78539 USA

来源：

JOURNAL OF ADVANCED COMPUTATIONAL INTELLIGENCE AND INTELLIGENT INFORMATICS | 2010年 / 14卷 / 03期

关键词：

formal concept analysis; human-machine interface; lattice structure; image processing; visualization;

D O I：

10.20965/jaciii.2010.p0303

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

In order to perform an interactive data-mining for huge image databases efficiently, a visualization interface based on Formal Concept Analysis (FCA) is proposed. The proposed interface system provides an intuitive lattice structure enabling users freely and easily to select FCA attributes and to view different aspects of the Hasse diagram of the lattice of a given image database. The investigation environment is implemented using C++ and the OpenCV library on a personal computer (CPU = 2.13 GHz, MM = 2 GB). In visualization experiments using 1,000 Corel Image Gallery images, we test image features such as color, edge, and face detectors as FCA attributes. Experimental analysis confirms the effectiveness of the proposed interface and its potential as an efficient data-mining tool.

引用

页码：303 / 308

页数：6

共 7 条

[1]

Bradski G, 2008, LEARNING OPENCV

[2]

CANNY JF, 1986, PAMI, V8, P6, DOI DOI 10.1109/TPAMI.1986.4767851

[3]

Davey B. A., 2002, INTRO LATTICES ORDER, DOI DOI 10.1017/CBO9780511809088

[4]

Feldman, 1973, PATTERN CLASSIFICATI

[5]

Ganter B., 1996, FORMAL CONCEPT ANAL

[6]

Marr D., 1982, VISION

[7]

Sawase K., 2009, J JAPAN SOC FUZZY TH, V21, P32

← 1 →