Human face comparison method, apparatus and system

Abstract

The invention provides a human face comparison method, apparatus and system. Multiple classification libraries are preset in a human face database; each classification library comprises multiple sub-libraries; and the classification libraries and / or the sub-libraries are configured with priorities. The method comprises the steps of receiving a comparison range, wherein the comparison range comprises at least one sub-library or classification library in the human face database, selected by a user; according to the comparison range and current calculation resources, determining a human face comparison mode, wherein the human face comparison mode includes parallel comparison or sequential traversal comparison; and according to the priorities and the determined human face comparison mode, performing human face comparison on a currently collected image in the comparison range to obtain a comparison result. The human face database adopted in the method comprises the classification librariesand the sub-libraries; during the human face comparison, the comparison range can be dynamically selected according to a user demand; and the reasonable comparison mode can be selected according to the comparison range and the current calculation resources, so that the human face comparison efficiency is improved while the deployment control scale meets the user demand.

Description

technical field [0001] The present invention relates to the technical field of face recognition, in particular to a face comparison method, device and system. Background technique [0002] Face recognition technology has been deployed and applied in many application fields. In the real application of dynamic face control such as public security, the face database needs to include not only the face database of key monitoring objects in the area under its jurisdiction, but also the entire province Even key wanted criminals across the country, as well as dangerous elements involved in terrorism and violence are added to the automatic identification comparison database, and the blacklisted face databases under control are basically at the million level. Reducing the size of the face database for deployment and control will lead to the problem of incomplete deployment and control. [0003] At present, many Internet companies use the method of adding hardware services and cloud c...

AI Technical Summary

Problems solved by technology

[0002] Face recognition technology has been deployed and applied in many application fields. In the real application of dynamic face control such as public security, the face database needs to include not only the face database of key monitoring objects in the area under its jurisdiction, but also the entire province Even key wanted criminals across the country, as well as dangerous elements involved in terrorism and violence are added to the automatic identification comparison database, and the blacklisted face databases under control are basically at the million level. Reducing the size of the face database for deployment and control will lead to the problem of incomplete deployment and control

[0003] At present, many Internet companies use the method of adding hardware services and cloud clusters to solve the above contradictions, but this method has a series of problems such as expensive construction costs, untimely data transmission, insufficient data confidentiality, high technical requirements for operation and maintenance, and difficult management. practical application problems, resulting in poor practical applicability

[0004] Aiming at the contradiction between the deployment scale and the recognition response speed mentioned above, it is difficult to meet the needs of users, and no effective solution has been proposed

Images