摘要
近年来,人工智能高速发展,各种神经网络层出不穷。介绍了生成式对抗网络的基本原理,以及GAN在图像、音频、视频3大领域中的应用,包括风格转换、音频生成、视频生成等。根据现有研究,分析了当前生成式对抗网络发展中的不足,并对未来的研究方向进行了展望。
In recent years,artificial intelligence has developed rapidly.Various neural networks have emerged endlessly.This paper introduces the basic principles of generative adversarial network,as well as GAN’s application in the three areas of image,audio,video,including image style transfer,audio generation,video generation,etc.The deficiencies of generative adversarial network and the future research directions are analyzed according to the existing research.
作者
才智
魏为民
孟繁星
刘畅
CAI Zhi;WEI Weimin;MENG Fanxing;LIU Chang(School of Computer Science and Technology,Shanghai University of Electric Power,Shanghai 200090,China)
出处
《上海电力大学学报》
CAS
2022年第2期199-202,共4页
Journal of Shanghai University of Electric Power
基金
上海市自然科学基金(16ZR1413100)。
关键词
神经网络
生成式对抗网络
图像处理
音频处理
视频处理
neural network
generative adversarial network
image processing
audio processing
video processing