Automatic image annotation algorithm based on the region of interest of the image
For improving the accuracy of image automatic annotation, the paper submits a new automatic annotation algorithm based on the region of interest ( ROI) of the image. First, it extracts the region of interest of the image by using the Maxshift algorithm of JPEG 2000, thereby establishing a ROI library, and then extracts the feature vector by using SIFT algorithm in the image library, and SVM is adopted to classify and annotate images automatically. Experimental results on images from Corel database show that this algorithm can improve the annotation accuracy.
automatic image annotationMaxshift algorithmSIFTk - Means clusterSVM
国家科技期刊平台
NSTL
万方数据
维普
