Automatic Image Annotation Based on Multi-scale Salient Region
Automatic image annotation is a challenging problem in image understanding areas. The existing models directly extract visual features from segmented image regions. Since segmented image regions may still have multi-objects, the extractive visual features may not effectively describe corresponding regions. In order to overcome the above problems, an image annotation model based on multi-scale salient region is proposed. In this model, first, each image is segmented by using multi-scale grid-based segmentation method. Second, global contrast-based method is used to extract the saliency maps from each image region. Third, visual features are extracted from each salient region. Finally, multi-scale visual features of image regions are fused and applied to automatic image annotation. Our model can improve the object descriptions of images and image regions. Experimental results conducted on Corel 5K datasets verify the effectiveness of proposed model.
- 1.Feng SL, Manmatha R, Lavrenko V (2004) Multiple Bernoulli relevance models for image and video annotation. In: Proceedings of the IEEE computer society conference on computer vision and pattern recognition, Washington, DC, pp 1002–1009Google Scholar
- 2.Jeon J, Lavrenko V, Manmatha R (2003) Automatic image annotation and retrieval using cross-media relevance models. In: Proceedings of the 26th annual international ACM SIGIR, Toronto, Canada, pp 119–126Google Scholar
- 6.Ming-Ming C, Guo-Xin Z, Mitra N, et al. (2011) Global contrast based salient region detection. In: Proceedings of the IEEE computer society conference on computer vision and pattern recognition, pp 409–416Google Scholar