Online Social Network Image Classification and Application Based on Deep Learning

Abstract

Apply deep belief networks to online social network image classification, taking Sina microblog as an example of online social network. Images and text information are acquired from Sina microblog developer platform. Deep belief networks use unsupervised learning algorithm learning image characteristics, which is applied to online social networking image classification tasks. In this paper, image classification which combines text features and content features obtained from the deep learning method is proposed. The data sets obtained from Sina microblog were used to evaluation. Finally this method is applied to the sports brand logo image classification, and its practical significance is to compare the attention of different sports brand in different times in different areas.