A novel method of automatic image annotation

Abstract

Automatic image annotation can improve the performance of image retrieval. Some methods of annotation have been proposed in the past years. In this paper, we introduce a novel annotation method based on non-linear regression model in order to annotate image accurately. Both the visual and the textual modalities are efficiently represented by a continuous feature vector, and are named by the visual blob vector and the semantic description vector, respectively. The task of annotation is to fit a rigorous mapping construction between the visual blob vectors and the semantic description vectors using a method based on least squares estimation. The advantages of the proposed method are conceptually simple, computationally efficient, scalable for huge amount of images and no priori knowledge of images and keywords. With a highly accurate approximation function, the experimental results demonstrate the improvement of annotation performance.