论文部分内容阅读
端到端的联合判别和生成式对抗神经网络(End-to-End Joint Discriminative and Generative Learning NeuralNetral,DG-Net),将生成对抗网络(Generative Adversarial Networks,GAN)和残差神经网络(Residual Network,Res Net)结合在一起,使行人重识别(Person Re-Identification,Re-ID)能够更好地利用数据生成模型来训练网络,实现数据生成模型和Re-ID网络