Dangerous goods identification algorithm based on Generative Adversarial Networks

Abstract

With the continuous progress of deep learning, the application has been extended to the identification of dangerous goods in the security inspection system. The quality of the identification algorithm directly determines the quality of the security inspection system. In this paper, a recognition algorithm based on Generative Adversarial Networks is proposed, which can continuously train the generation model and the discrimination model, and collect the image data samples that are blocked or too small, and then generate images with high similarity, so as to achieve the purpose of detecting and tracking the hidden or small dangerous goods. The experimental results show that the algorithm can effectively detect the dangerous goods and greatly improve the tracking effect.