Image Semantic Annotation Based on Gaussian Mixture Model
Kong Fanhui · 2011
Automatic semantic annotation of an image is very important and very challenging in content-based image retrieval. Color and texture features are integrated to describe the low-level characters of images, and GMM is used for semantic annotation of images. Experimental results show that multi-feature image retrieval is more effective than a single feature, and integration of multiple features with GMM can be successfully used for image semantic annotation. In an image library database that contains 1000 images the experimental results show that the proposed method has better performance.