Biological characteristic management method, system and device and medium

Abstract

The invention provides a biological feature management method, system and device and a medium. The method comprises the following steps: converting biological feature vectors extracted from a plurality of biological models of the same category; and analyzing the obtained biological characteristics based on the converted biological characteristic vector to obtain an analysis result. According to the method, the biological feature vectors are converted, so that the biological models can be compatible during common use, and the compatibility between the biological models is improved; and analyzing the acquired biometric feature vector based on the converted biometric feature vector to obtain an analysis result. According to the method, the biological feature vectors are converted, so that thebiological feature vectors extracted from different biological models can be fused, and some biological models do not need to collect part of the biological feature vectors again, therefore, the storage space required for storing the biological feature vector can be saved, and the time required for re-extracting the biological feature vector can be reduced, so that the biological model can be updated smoothly in real time.

Description

technical field [0001] The present invention relates to the technical field of biological identification, in particular to a biological feature management method, system, equipment and medium. Background technique [0002] Face comparison is a hot field of current AI applications, and the similarity between faces is reflected in the similarity between the feature vectors corresponding to faces. In this field, regardless of face comparison, face retrieval, face clustering and other scenarios, a large number of vector calculations are required. In actual scenarios, in systems such as face recognition and comparison, it is first necessary to extract face feature vectors, and then perform corresponding vector calculations. The face model will continue to iterate with the update of the training data and the optimization of the training algorithm; at the same time, some scenarios will also encounter a variety of face models provided by different institutions, companies, and scien...

AI Technical Summary

Problems solved by technology

Once a variety of face models are used in the same project, whether it is a new replacement or multiple models used in parallel, it is difficult to mix and calculate multi-source face feature vectors, that is, multi-source face feature vectors cannot be fused and compared

Images