Multi-modal composite coding image retrieval method and system

Abstract

The invention provides a multi-modal composite coding image retrieval method and system, and the method comprises the steps: obtaining to-be-retrieved information, and judging the data type of the to-be-retrieved information; selecting the coarse-grained retrieval or fine-grained retrieval according to the data type of the to-be-retrieved information, wherein the fine-grained retrieval comprises the steps that when the input to-be-retrieved information comprises image description data and text description data at the same time, the two features are fused to obtain composite features, and then data retrieval is conducted through the composite features. According to the method, coarse-grained retrieval or fine-grained retrieval can be selected according to the data type, the composite features can be obtained by fusing the two features, and then data retrieval of fine-grained retrieval is carried out through the composite features; according to the method, the database construction efficiency can be improved, the flexibility of a retrieval mode is improved, the retrieval difficulty is reduced, and a solution which is uniform in precision and generalization performance and more flexible in switching between fuzzy retrieval and accurate retrieval is provided for image retrieval.

Description

technical field [0001] The invention relates to the field of computer applications, in particular to a method and system for retrieving multimodal composite coded images. Background technique [0002] Image retrieval technology has been widely used in various fields, such as product search, video content understanding, etc. However, due to the diversity of images and the complexity of image content, the research on efficient and accurate image retrieval methods has always been an important research topic in the field of machine vision. [0003] At present, the technical process of image retrieval mainly consists of three steps: feature extraction, feature encoding and database indexing. According to different characteristics, image retrieval is generally divided into text-based retrieval and image content-based retrieval, and these two technologies have also been widely used and researched. Text-based retrieval generally marks the attributes of each dimension of the image,...

AI Technical Summary

Problems solved by technology

Text-based retrieval generally marks the attributes of each dimension of the image, and only images with the same attributes can be matched during retrieval. The accuracy is high but the generalization performance is poor, and a large number of images need to be manually annotated in advance, and the time cost is relatively high. high

Retrieval based on image content uses image semantics as clues for processing, completes the extraction of low- and high-dimensional features of images based on image processing technology, and performs matching, which can retrieve images with the same or similar characteristics, and has certain generalization performance , less manual participation; but based on the diversity and complexity of image features, there may be a large number of mismatches, for example, it is difficult to distinguish when there are images of the same category but not the same target image

In addition, the text description of the image content is completed by annotating the text semantics of the image in advance, which is expensive to apply and difficult to maintain

Images