Maximum A Posteriori Estimation of Distances Between Deep Features in Still-to-Video Face Recognition
This addresses face recognition in video for security or surveillance applications, but appears incremental as it builds on existing deep features and dissimilarity measures.
The paper tackles the still-to-video face recognition problem for small sample sizes by proposing a Maximum A Posteriori estimation method that maximizes the joint probabilistic density of distances between deep features, achieving 3-5% higher accuracy compared to conventional frame aggregation methods.
The paper deals with the still-to-video face recognition for the small sample size problem based on computation of distances between high-dimensional deep bottleneck features. We present the novel statistical recognition method, in which the still-to-video recognition task is casted into Maximum A Posteriori estimation. In this method we maximize the joint probabilistic density of the distances to all reference still images. It is shown that this likelihood can be estimated with the known asymptotically normal distribution of the Kullback-Leibler discriminations between nonnegative features. The experimental study with the LFW (Labeled Faces in the Wild), YTF (YouTube Faces) and IJB-A (IARPA Janus Benchmark A) datasets has been provided. We demonstrated, that the proposed approach can be applied with the state-of-the-art deep features and dissimilarity measures. Our algorithm achieves 3-5% higher accuracy when compared with conventional aggregation of decisions obtained for all frames.