生成对抗网络在计算机视觉领域的应用  

The Applications of Generative Adversarial Networks in the Field of Computer Vision

作　　者：江春雨 程琳[1] 黎晓明亮 

机构地区：[1]兰州大学信息科学与工程学院,甘肃兰州

出　　处：《计算机科学与应用》2018年第11期1726-1733,共8页Computer Science and Application

摘　　要：生成对抗网络(GAN, Generative Adversarial Networks)的出现是计算机视觉领域又一里程碑式的发展,它为解决各种图像预测问题提供了新型工具。以此为目的,本文通过相关文献的调研以及结合最新成果的研究,突出了生成对抗网络的无监督学习算法特征。目前生成对抗网络拓扑结构的研究在不断优化着基础生成对抗网络的性能;在改进难收敛、模式崩溃等缺点方面发挥着各自的长处,已然成为当下研究的热点方向之一。The emergence of generative adversarial networks is another milestone in the field of computer vision. It provides a new tool for solving various image prediction problems. For this purpose, from the investigation of relevant literatures and latest research results, the characteristics of unsupervised learning algorithms are highlighted. At present, the topology of the generative adversarial networks is continuously optimized to its basic performance. It has become one of the hot topics of current research to play their respective strengths in improving the difficulties such as non-convergence, mode collapse and so on.

关 键 词：生成对抗网络 计算机视觉 图像预测 无监督学习 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]