Semantic propagation and mixed multi-instance learning-based Web image retrieval method

Abstract

The invention belongs to the technical field of image processing and particularly provides a semantic propagation and mixed multi-instance learning-based Web image retrieval method. Web image retrieval is performed by combining visual characteristics of images with text information. The method comprises the steps of representing the images as BoW models first, then clustering the images according to visual similarity and text similarity, and propagating semantic characteristics of the images into visual eigenvectors of the images through universal visual vocabularies in a text class; and in a related feedback stage, introducing a mixed multi-instance learning algorithm, thereby solving the small sample problem in an actual retrieval process. Compared with a conventional CBIR (Content Based Image Retrieval) frame, the retrieval method has the advantages that the semantic characteristics of the images are propagated to the visual characteristics by utilizing the text information of the internet images in a cross-modal mode, and semi-supervised learning is introduced in related feedback based on multi-instance learning to cope with the small sample problem, so that a semantic gap can be effectively reduced and the Web image retrieval performance can be improved.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to a Web image retrieval method based on semantic propagation and mixed multi-instance learning. Background technique [0002] In the network environment, images are generally embedded in Web pages and published with rich text information, such as tags, file names, URL information, and image context. For Web image retrieval, TBIR (Text-based Image Retrieval) based on text information and CBlR (Content-based Image Retrieval) based on image visual features have their own advantages and disadvantages. To a certain extent, TBIR avoids the problem of identifying complex visual elements, makes full use of web page context and hypertext structure information, and conforms to people's familiar retrieval habits, and is simple to implement. However, because it is still limited to the scope of text retrieval, it uses controlled vocabulary To describe the image, it is pron...

AI Technical Summary

Problems solved by technology

However, this method will generate a very large and complex relationship network in the process of relationship propagation, and the calculation is complex; moreover, the propagation process will generate a large number of auxiliary visual vocabulary, thereby reducing the accuracy of image retrieval

Images