Artificial Intelligence (AI) and Machine Learning (ML) in Biometric Data Fusion

This State-of-the-Art Report focuses on recent advances in biometric recognition technologies and the application of artificial intelligence and machine learning to recognition tasks in multimodal identification systems. Multimodal systems use “feature-level” data fusion (e.g., periocular and gait recognition), which provides faster reference set retrieval across identity templates and significantly improves recognition accuracy over a unimodal system. Specifically, the recent field of biometric data fusion holds promise to deliver improved biometric data sample capture  and analysis to the warfighter regardless of disguised, altered, or occluded facial characteristics. The use of convolutional neural networks, deep neural networks, and recurrent neural networks in biometric data fusion is also explained in this report. In addition, topics in leading-edge biometric recognition research are also presented.