摘要

由于受到光照和成像设备等条件因素的影响,采集到的单帧人脸图像分辨率低,无法进行准确人脸识别,所以需要图像超分辨率重建。而利用SRGAN模型在进行人脸超分辨率重建过程中,易出现梯度消失或爆炸的问题,严重影响了重建图像的精度和质量。针对上述问题,提出了基于生成对抗网络的改进人脸超分辨率重建算法,在SRGAN结合WGA-N的基础上引入Wasserstein散度,并将其最大化得到最优化标量函数T,实现去掉Lipschit-z约束能够直接得到Wassertein距离,并通过最小化Wassertein距离得到生成网络的目标函数,最终改进的模型提高了重建图像的质量。实验结果表明,该方法能够生成高分辨率的人脸图像,在主观视觉和客观评价指标均同比优于DRCN、FSRCNN、SRGANWGAN、VDSR和DRRN模型。

访问全文
Internet of Things Technology and Value Added
Wortmann Felix*; Fluechter Kristina
Business & Information Systems Engineering, 2015, 57(3): 221-224.

512

被引

访问全文
Smart logistics based on the internet of things technology: an overview
Ding Yangke; Jin Mingzhou; Li Sen; Feng Dingzhong
International Journal of Logistics-Research and Applications, 2020, 24(4): 1-23.

193

被引

Application of the Internet of Things Technology in Smart Grid
Dong Cuiying*; Zhou Jianli; Wang Moqi
International Conference on Social Science, Education Management and Sports Education (SSEMSE), 2015-04-10 to 2015-04-11.
访问全文
Power internet of things technology with energy and information fusion
Chen, Haoyong*; Cai, Weijun; Chen, Jianrun
Power System Protection and Control, 2021, 49(22): 8-17.

10

被引

Research Status and Development Trend of Internet of Things Technology
Yan Qian*
International Conference on Intelligent Systems Research and Mechatronics Engineering (ISRME), 2015-04-11 to 2015-04-13.
访问全文
How the internet of things technology improves agricultural efficiency
Duguma, Amenu Leta; Bai, Xiuguang*
Artificial Intelligence Review, 2024, 58(2): 63.
Development and application of Internet of things technology in Greenhouse
Gong He; Ju Hongyan*; Zhao Zhongwei; Liu Bin
4th International Conference on Machinery, Materials and Information Technology Applications (ICMMITA), 2016-12-10 to 2016-12-11.
The Application of the Internet of Things Technology on the People Counting
Ren Tong Xuan*
International Conference on Electrical Engineering and Mechanical Automation (ICEEMA), 2015-06-13 to 2015-06-14.