Image annotation via graph learning
| Autor: | Mingjing Li, Qingshan Liu, Hanqing Lu, Jing Liu, Songde Ma |
|---|---|
| Rok vydání: | 2009 |
| Předmět: |
Training set
Information retrieval ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION Computer Science::Computation and Language (Computational Linguistics and Natural Language and Speech Processing) Image processing Similitude Annotation Automatic image annotation Artificial Intelligence Computer Science::Computer Vision and Pattern Recognition Signal Processing Graph (abstract data type) Pairwise comparison Computer Vision and Pattern Recognition Image retrieval Software Mathematics |
| Zdroj: | Pattern Recognition. 42:218-228 |
| ISSN: | 0031-3203 |
| Popis: | Image annotation has been an active research topic in recent years due to its potential impact on both image understanding and web image search. In this paper, we propose a graph learning framework for image annotation. First, the image-based graph learning is performed to obtain the candidate annotations for each image. In order to capture the complex distribution of image data, we propose a Nearest Spanning Chain (NSC) method to construct the image-based graph, whose edge-weights are derived from the chain-wise statistical information instead of the traditional pairwise similarities. Second, the word-based graph learning is developed to refine the relationships between images and words to get final annotations for each image. To enrich the representation of the word-based graph, we design two types of word correlations based on web search results besides the word co-occurrence in the training set. The effectiveness of the proposed solution is demonstrated from the experiments on the Corel dataset and a web image dataset. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect