| 摘要: |
| 图像中存在的纹理、颜色和形状等异构视觉特征,在表示特定高层语义时所起作用的重要程度不同.为了在图像标注过程中更加有效地利用这些异构特征,提出了一种基于组稀疏(group sparsity)的多核学习方法(multiplekernel learning with group sparsity,简称MKLGS),为不同图像语义选择不同的组群特征.MKLGS 先将包含多种异构特征的非线性图像数据映射到一个希尔伯特空间,然后利用希尔伯特空间中的核函数以及组LASSO(groupLASSO)对每个图像类别选择最具区别性特征的集合,最终训练得到分类模型对图像进行标注.通过与目前其他图像标注算法进行对比,实验结果表明,基于组稀疏的多核学习方法在图像标注中能取得很好的效果. |
| 关键词: 组LASSO 组稀疏 多核学习 特征选择 图像标注 |
| DOI：10.3724/SP.J.1001.2012.04154 |
| 分类号: |
| 基金项目:国家自然科学基金(61070068, 60833006); 国家重点基础研究发展计划(973)(2010CB327904); 核高基项目(2010ZX01042-002-003) |
|
| Image Annotation by the Multiple Kernel Learning with Group Sparsity Effect |
|
YUAN Ying, SHAO Jian, WU Fei, ZHUANG Yue-Ting
|
|
Institute of Artificial Intelligence, College of Computer Science and Technology, Zhejiang University, Hangzhou 310027, China
|
| Abstract: |
| Since different kinds of heterogeneous features (such as color, shape and texture) in image shave different intrinsic discriminative power for image understanding, this paper proposes a multiple kernel learning with group sparsity (MKLGS) to select groups of discriminative features for image annotation to effectively utilize those heterogeneous visual features. Given each image label, the MKLGS method embeds the nonlinearity image data with discriminative features into a Hilbert space, and then utilizes the kernel function in the Hilbert space and group LASSO to select groups of discriminative features. Finally, a classification model can be trained for image annotation. In comparison to other image annotation algorithms, experiments show that the proposed MKLGS for imageannotation achieves a better performance. |
| Key words: group LASSO group sparsity multiple kernel learning feature selection image annotation |
