Adaptive Content-based Image Retrieval using Semi-Supervised Learning Method

使用半监督学习方法的自适应基于内容的图像检索

• 批准号: 22500152
• 负责人: MATSUMOTO TETSUYA
• 金额: $ 1.83万
• 依托单位: Nagoya University
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Scientific Research (C)
• 财政年份: 2010
• 资助国家: 日本
• 起止时间: 2010-10-20 至 2013-03-31
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-22500152/
• 关键词: 画像情報処理; 類似画像検索; 部分教師付学習; 適応的画像内容検索

Automatic image annotation is a hopeful sub-technique for image database retrieval. We have been constructing a generative model system for automatic image annotation using semi-supervised learning method. As it can be easily unstable for the higher dimensions, we must apply a dimensionality reduction method in advance. Generally, conventional supervised dimensionality reduction method (using labeled samples) suffers from the degenerate covariance matrix problem in the case of a small number of samples. On the other hand, unsupervised dimensionality reduction method (using unlabeled samples) can't recognize the differences among the categories properly. In this study, we propose a novel semi-supervised dimensionality reduction method using a small number of labeled samples and a large number of unlabeled samples. By the result of experiments, the classification rate of the proposed method was 5.1 points better than that of the unsupervised method.

图像自动标注是图像数据库检索中一项很有希望的子技术。我们一直在使用半监督学习方法构建一个自动图像标注的生成模型系统。由于它在高维数时很容易不稳定，我们必须提前应用降维方法。通常，传统的有监督降维方法（使用标记样本）在样本数量较少的情况下会遇到退化协方差矩阵问题。另一方面，无监督的降维方法（使用未标记的样本）不能正确识别类别之间的差异。在这项研究中，我们提出了一种新的半监督降维方法，使用少量的标记样本和大量的未标记样本。实验结果表明，该方法的分类正确率比无监督方法提高了5.1个百分点。

项目成果

Incorporating Contextual Information into Bag-of-Visual-Words Framework for Effective Object Categorization

• DOI: 10.1587/transinf.e95.d.3060
• 发表时间: 2012-12
• 期刊: IEICE Trans. Inf. Syst.
• 作者: Shuang Bai;Tetsuya Matsumoto;Y. Takeuchi;H. Kudo;N. Ohnishi

A Novel Method for Exploring Patch-level Context to Improve Image Categorization Performance

• DOI: 10.1541/ieejeiss.133.2264
• 发表时间: 2013-12
• 期刊: Ieej Transactions on Electronics, Information and Systems
• 作者: Shuang Bai;Tetsuya Matsumoto;Y. Takeuchi;H. Kudo;N. Ohnishi

Informative patches sampling for image classification by utilizing bottom-up and top-down information

• DOI: 10.1007/s00138-012-0473-x
• 发表时间: 2012-02
• 期刊: Machine Vision and Applications
• 影响因子: 3.3
• 作者: Shuang Bai;Tetsuya Matsumoto;Y. Takeuchi;H. Kudo;N. Ohnishi

Semi-supervised Dimensionality Reduction for Content Based Image Retrieval

基于内容的图像检索的半监督降维

• 发表时间: 2011
• 作者: T. Matsumoto;M. Yoshida;N. Ohnishi
• 通讯作者: N. Ohnishi

Scene Classification based on Category-Specific Representations Created through Prototype Future Selection

基于通过原型未来选择创建的特定类别表示的场景分类

• 发表时间: 2012
• 作者: S. Bai;H. Kudo;Y. Takeuchi;T. Matsumoto;N. Ohnishi
• 通讯作者: N. Ohnishi