High-hardware elasticity accurate face sampling and recognition method

Abstract

The invention discloses a high-hardware elasticity accurate face sampling and recognition method, which comprises a face sampling database building method and a face recognition method. Through introducing "Internet +" and "cloud computing" concepts, unaware separation between a front end and a rear end is realized, and a queue thinking is delicately used to realize double transparent proxy between "field device" and "high-speed computing cloud". The cloud thinking is used to deploy centralized computing capabilities, multiple low-end devices are allowed to share one computing device at the same time, the cost is controlled, waste of resources is also avoided, the advantage of the high-hardware elasticity structure also lies in that a client can finally choose to be on-cloud or not on-cloud, and a possibility is provided for the optimal balance between application and the cost.

Description

technical field [0001] The mechanism of the invention relates to a precise face sampling and recognition method, specifically a high hardware flexibility precise face sampling and recognition method, which belongs to the technical field of information collection and recognition. Background technique [0002] Most of the existing common face recognition technologies are concentrated in the conventional attendance and security fields, and system integrators provide hardware-based implementation solutions. Due to the limited computing power of the equipment, it can only collect very limited plane face information, so It is not compatible with common situations such as face rotation, wearing accessories, and hairstyle changes, and the user experience is poor. [0003] There are a small number of enterprises that can accurately recognize faces, and can achieve better compatibility and user experience by extracting more facial information. However, with the huge increase in data, ...

AI Technical Summary

Problems solved by technology

[0002] Most of the existing common face recognition technologies are concentrated in the conventional attendance and security fields, and system integrators provide hardware-based implementation solutions. Due to the limited computing power of the equipment, it can only collect very limited plane face information, so It is not compatible with common situations such as face rotation, wearing accessories, and hairstyle changes, and the user experience is poor

[0003] There are a small number of enterprises that can accurately recognize faces, and can achieve better compatibility and user experience by extracting more facial information. However, with the huge increase in data, related recognition and Matching calculations cannot be completed by conventional computers, so it is not conducive to the rollout of conventional applications

Moreover, when there are multiple concurrent requests on the front-end equipment, it is very easy to be blocked or delayed; when the type of front-end equipment or the demand of the scene changes, automatic intelligent adaptation cannot be performed, and frequent reconfiguration of the back-end processing platform is required.

Images