A mobile-cloud architecture provides a practical platform for performing face recognition on a mobile device. Firstly, even though CPU, memory in mobile phones have been greatly improved, resources demanding applications like face recognition still cannot run on mobile phones. Besides, if send face images to clouds, network latency will not be accepted. However, cloudlet provides an accelerator for face recognition and lessens compute burden on mobile phones which have limited computational capabilities and battery life.
I、Background
人脸检测和识别在日常生活中的应用日渐广泛,例如:安全监控,人脸识别登录,刷脸支付等。随着智能手机的发展,手机APP成为了人脸识别使用最广泛的地方。在过去的十几年间,智能手机经历了一段飞速发展的时期,触屏显示器的使用、从单核处理器到多和处理器、存储从几百M扩展到几百G,这些技术的提高和进步使得智能手机可以运行更加复杂的应用。然而即使如此这样,智能手机的计算和存储能力仍然有限,再加上有限的电池电量,所以一些计算密集型的应用仍然不能在手机上运行,例如:实施人脸检测和识别。其中一种解决方案就是基于云端的方法,将