Meta-Learning for Finger Vein Recognition in Internet of Things Smart Home Security

Abstract

Recently, convolutional neural networks for finger vein recognition have gained attention, but their application in IoT smart home security is underexplored. Existing methods typically require networks to identify all categories in a dataset, leading to high parameter demands, which is inefficient given the small, dynamic user groups (3-5 users) in smart homes. To address this, we propose a finger vein recognition system based on meta-learning. Our approach frames recognition as a meta-learning task, introducing a dynamic, exponentially-weighted multistep loss optimization to enhance the model-agnostic meta-learning process. This allows quick adaptation to new tasks with minimal data. Additionally, we design an adaptive recognition scheme that updates network parameters without altering the structure for various users. Experiments on public datasets confirm the effectiveness of our system in IoT smart home security, achieving excellent recognition performance.