Automatic region-based image annotation using an improved multiple-instance learning algorithm
文献类型：期刊论文
| 作者 | Songhe Feng; De Xu ; Bing Li
|
| 刊名 | Chinese Journal of Electronics
 |
| 出版日期 | 2008 |
| 卷号 | 17期号:1页码:43-47 |
| 关键词 | Image Annotation |
| 英文摘要 | Many existing image annotation algorithms work under probabilistic modeling mechanism. In this paper, we formulate the problem as a variation of supervised learning task and propose an Improved CitationkNN (ICKNN) Multiple-instance learning (MIL) algorithm for automatic image annotation. In contrast with the existing MIL based image annotation algorithm which intends to learn an explicit correspondence between image regions and keywords, here we annotate the keywords on the entire image instead of its regions. Concretely, we first explore the concept of Confidence weight (CW) for every training bag (image) to reflect the relevance extent between a bag and a semantic keyword. It can be treated as a stage of re-ranking on training set before annotation starts. Moreover, a modified hausdorff distance is adopted for the ICKNN algorithm to solve the automatic annotation problem. The proposed annotation approach demonstrates a promising performance over 5,000 images from COREL dataset, as compared with some current algorithms in the literature. |
| 源URL | [http://ir.ia.ac.cn/handle/173211/20391]  |
| 专题 | 自动化研究所_09年以前成果 |
| 作者单位 | Beijing Jiaotong University |
| 推荐引用方式 GB/T 7714 | Songhe Feng,De Xu,Bing Li. Automatic region-based image annotation using an improved multiple-instance learning algorithm[J]. Chinese Journal of Electronics,2008,17(1):43-47. |
| APA | Songhe Feng,De Xu,&Bing Li.(2008).Automatic region-based image annotation using an improved multiple-instance learning algorithm.Chinese Journal of Electronics,17(1),43-47. |
| MLA | Songhe Feng,et al."Automatic region-based image annotation using an improved multiple-instance learning algorithm".Chinese Journal of Electronics 17.1(2008):43-47. |
入库方式： OAI收割
来源：自动化研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。