基于生成对抗网络的通信网络数据传输加密技术  

作　　者：邱馨亚 

机构地区：[1]长江职业学院机电汽车学院,湖北武汉430074

出　　处：《现代传输》2025年第1期48-51,共4页Modern Transmission

摘　　要：由于网络数据传输加密与解密的过程中常存在异常交叉,导致数据的有效传输占比难以得到保障。为此,本研究提出基于生成对抗网络的通信网络数据传输加密技术。首先,建立生成对抗网络模型,在训练模型后得到交叉熵收敛的无嵌入隐写生成器,再将原始数据信息和噪声向量输入到生成器中,生成对应的隐写数据信息。在通信网络的接收端,利用包含四个反卷积层的生成器对隐写数据进行提取,获取原始的数据信息。测试结果表明:应用该方法后,数据的有效传输占比始终稳定在98.0%以上,最大值达到了99.0%。Due to the abnormal crossover in the process of network data transmission encryption and decryption,it is difficult to guarantee the effective proportion of data transmission.Therefore,this study proposes the data transmission encryption technology of communication network based on generated adversarial network.First,the generative adversarial network model is established,and the cross-entropy convergence is obtained after training the model,and then the raw data information and noise vector are input into the generator to generate the corresponding stepwrite data information.At the receiving end of the communication network,the generator containing four deconvolution layers is used to extract the implicit data and obtain the raw data information.The test results show that:after the application of this method,the effective transmission ratio of data is always stable above 98.0%,and the maximum value reaches 99.0%.

关 键 词：生成对抗网络 通信网络数据 传输加密技术 交叉熵 无嵌入隐写生成器 隐写数据信息 反卷积层 生成器 

分 类 号：TP309.7[自动化与计算机技术—计算机系统结构]