Research on Text Detection in Network Advertisement Picture Based on Depth Learning

Abstract

with the development of network information technology, people use more and more popular network technology. Online advertising behavior is also more and more; especially the text existing in the way needs to spread the picture. In order to quickly detect the sensitive words in the picture text, the introduction of the depth learning methods is used for text detection and recognition. RBM is used to judge and select the text of the advertisement image area. The BP neural network algorithm and the deep confidence network (DBN) algorithm are used to detect the sensitive information corresponding to the text. Sensitive text detection theory analysis and experimental data show that the algorithm has low complexity and fast detection speed.