Image Annotation Based on Color and Texture features

This paper presents a method for the establishment of mapping between low-level visual features and high-level semantic features based on Gaussian mixture model (GMM). We combine color features and texture features based on Gabor filter, an image annotation algorithm based on GMM clustering is put forward through the research in image processing. In an image database that contains 1000 images, the experimental results show that the proposed method has better performance.