A GAN-Based Image Generation Method for X-Ray Security Prohibited Items
Recognizing prohibited items intelligently is significant for automatic X-ray baggage security screening.In this field,Convolutional Neural Network(CNN)based methods are more attractive in X-ray image contents analysis.Since training a reliable CNN model for prohibited item detection traditionally requires large amounts of data,we propose a method of X-ray prohibited item image generation using recently presented Generative Adversarial Networks(GANs).First,a novel posebased classification method of items is presented to classify and label the training images.Then,the CT-GAN model is applied to generate many realistic images.To increase the diversity,we improve the CGAN model.Finally,a simple CNN model is employed to verify whether or not the generated images belong to the same item class as the training images.
Generative Adversarial Network X-ray prohibited item images Image generation Feature transformation
Zihao Zhao Haigang Zhang Jinfeng Yang
Tianjin Key Lab for Advanced Signal Processing,Civil Aviation University of China,Tianjin,China
国际会议
广州
英文
420-430
2018-11-23（万方平台首次上网日期，不代表论文的发表时间）