Finger vein recognition algorithm based on automatic enhancement contrast learning

• Main Authors: YANG KAIYUAN, LI YANTAO, QIN HUAFENG, LUO HUAXIU
• Format: Patent
• Language: Chinese; English
• Published: 25.06.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; PHYSICS

The invention provides a finger vein recognition algorithm based on automatic enhancement contrast learning, a discriminator and a distributor are respectively introduced based on an enhancement network in a contrast learning model without negative samples, the discriminator and the enhancement network are in a confrontation relationship, and the distributor is used for constraining the enhancement network to avoid generating a vein picture deviating from original data distribution. A better enhanced vein image sample pair is generated through automatic enhancement of adversarial learning, and the performance of a contrast learning model is improved. Compared with complex and tedious traditional manual selection of enhancement types and enhancement parameters, the method is more suitable for small data set scenes in various fields, and a better recognition effect is achieved on limited labeled data. 本发明提供了一种基于自动增强对比学习的指静脉识别算法，基于无负样本的对比学习模型中的增强网络，分别引入鉴别器和分布器，鉴别器和增强网络呈对抗关系，而分布器用于约束增强网络，避免产生偏离原数据分布的静脉图片。实现通过对抗学习