Method and device for automatic recognition of blood relationship based on face image analysis

Abstract

This application proposes a blood relationship automatic recognition method and device based on face image analysis, which relates to the technical field of biometric feature recognition, wherein the method includes: acquiring a pair of face images to be tested, and performing size measurement on the pair of face images to be tested Trim to obtain the processed image pair; load the trained model, input the processed image pair into the trained model, and output the shared feature vector and unique feature vector; calculate the cosine similarity of the shared feature vector, and then convert the cosine The similarity is compared with the optimal threshold to obtain the prediction result; the prediction result is stored in the database and displayed on the system interface. The face blood relationship recognition model proposed by this application using the above scheme can better extract the inherent genetic features of the face, improve the robustness and discrimination of face blood relationship recognition, and effectively improve the accuracy of model recognition.

Description

Technical field [0001] The present application relates to the field of biometric identification, and in particular, there is a method and apparatus for automatic recognition method and apparatus based on blood relationship based on face image analysis. Background technique [0002] Psychologists have found that human visual recognition has been recognized by human face photos. Enconspiled by this, the field of computer visual fields will begin to pay attention to the use of calculation methods to verify blood relationships. Human face blood relationship recognition has a wide range of application prospects in social production and life, including: disabled children and elderly search, family photo album assembly, sect research, social media analysis, and image labels. According to reports, China will have a small number of missing children and disappearing the elderly. Although the Ministry of Public Security implements a rapid search and DNA blood collection comparison mechanism...

AI Technical Summary

Problems solved by technology

[0007] For this reason, the first purpose of this application is to propose a blood relationship automatic recognition method based on face image analysis, which solves the problems caused by the practical difficulties and technical bottlenecks in the training data, pattern representation, recognition model and other aspects of the existing technology. The current recognition model still has deficiencies and defects in terms of discriminative power and robustness. The model is designed and optimized from two levels of feature extraction and recognition model, so that the model can better learn the inherent genetic feature representation of the face. Improve the robustness and discriminative power of face blood relationship recognition, effectively improve the accuracy of model recognition, and propose a method based on family label classification, which overcomes the problem of unbalanced sample distribution to a certain extent

Images