Method for automatically labeling images based on community potential subject excavation

Abstract

The invention discloses a method for automatically labeling images based on community potential subject excavation, which comprises the following steps: 1) adopting a hidden Dirichlet allocation modelto excavate implicit subjects in a single community; 2) after obtaining the probability distribution of image labels and the implicit subjects by analyzing community potential subjects, deleting theimage labels of which the probability of the community image labels and the implicit subjects is smaller than a set value k to perform 'de-noising' filtration on the community image labels; 3) generating image candidate labeling labels of the images to be labeled by propagating similar image labels; 4) optimizing the image candidate labeling labels according to the relativity between the image candidate labeling labels and the implicit subjects of the images; and 5) obtaining the final labeling result of the images through the information fusion of a plurality of communities. The method makesfull use of the information on different communities in which the images are positioned and the information on the community potential subjects in a social shared network to label the images, and compared with the conventional labeling method, the generated labeling result is more accurate.

Description

technical field [0001] The invention relates to the field of automatic labeling of images, in particular to a method for automatic labeling of images based on a social sharing network. Background technique [0002] We With the rapid development of network and multimedia technology, the number of images on the Internet is growing explosively. According to statistics, in 2008, Google has indexed 1 trillion web pages, including more than billions of image data. In recent years, the sharing network has attracted special attention from Internet users. On Flickr, a popular annotation website that provides digital image sharing, the number of images indexed has exceeded 3 billion, and it is growing rapidly at a rate of several million per month. [0003] The image tag information manually added by Internet users to Flickr images has brought great convenience to the efficient management and retrieval of images. However, through an in-depth analysis of the results of manual annotat...

AI Technical Summary

Problems solved by technology

[0007] The semantics of shared network images are complex. An image often contains multiple different subject information at the same time. For example, an image may contain not only subject information such as "Sky" and "Clouds", but also "Water" and "River". and other subject information

[0008] Due to the above-mentioned characteristics of shared network images, it is difficult to effectively label them using traditional algorithms

Images