Face feature updating method based on shared operation log

Abstract

The invention discloses a face feature updating method based on a shared operation log. The method comprises the steps of performing operation log recording on operation of a face library; when the video intelligent analysis application is started, importing all face features of a face library into a memory and storing the face features in a list form; when face comparison operation is involved, firstly, checking operation log records of a face feature library, when and only when new operation log information exists, updating the operation in a memory face list correspondingly, emptying the operation day, and then performing subsequent face comparison operation. According to the face feature updating method based on the shared operation log, the change condition of the face library in thedatabase can be updated into the face list in the memory in time, so that the change condition is kept consistent with the face library in the database, and normal operation of a face comparison application program cannot be influenced.

Description

technical field [0001] The invention relates to the technical field of face recognition, in particular to a face feature update method based on a shared operation log with good real-time performance and high face comparison accuracy. Background technique [0002] With the development of smart security, more and more intelligent applications are implemented in the actual construction of smart security, especially the application of face recognition, which plays an important role in the construction and development of smart communities, such as face access control and Applications such as face recognition check-in make the work of community security management more intelligent and digital. For people entering and leaving the community, through face recognition, accurate records are made. Opening the door with faces will also change the way people enter and exit. It has truly achieved Authentication of authority to enter and exit the community without media and instant detectio...

AI Technical Summary

Problems solved by technology

However, when performing face comparison search, there are often problems such as large amount of calculation and long waiting time. Therefore, it is proposed to preload the features of the face database into the memory to form a list, and use parallel computing methods such as GPU to perform face comparison. Action on search results

However, there may be additions, deletions, and changes to the face database at any time, and the modification does not happen from time to time. It has the characteristics of large randomness and long time intervals. How can the change operation of the face database be synchronized to the video intelligent analysis process? There is no corresponding better solution for the face list in memory

Images