科技创新与应用2026，Vol.16Issue(7)：18-21,4.DOI:10.19981/j.CN23-1581/G3.2026.07.004
亲子关系验证的端到端深度度量学习技术
摘要
Abstract
As an important application in the field of identification,parentage verification is in increasing demand in social life and legal affairs.Traditional biometric recognition methods often face challenges such as insufficient accuracy when dealing with complex facial similarities between family relatives.In recent years,deep learning,especially metric learning technologies,has shown strong potential in dealing with image similarity problems.Therefore,this paper discusses parent-child relationship verification technology based on end-to-end deep metric learning,and proposes a new model framework and optimization strategy,aiming to directly learn more discriminant parent-child feature representations and similarity measures from data.Criteria,in the expectation of significantly improving the accuracy and robustness of parent-child relationship discrimination,and providing more reliable technical support for related applications.关键词
亲子关系验证/端到端深度学习/度量学习/模型框架/优化策略Key words
parent-child relationship verification/end-to-end deep learning/metric learning/model framework/optimization strategies分类
信息技术与安全科学引用本文复制引用
兰天,杨伟樱,刘月,李晔..亲子关系验证的端到端深度度量学习技术[J].科技创新与应用,2026,16(7):18-21,4.基金项目
陕西省教育厅科学研究计划项目(24JK0320) （24JK0320）
