Attention-Aware Transformer-Based Aggregation Network for Video Periocular Recognition

TL;DR

This paper introduces an attention-aware transformer-based network for video periocular recognition, improving robustness and accuracy in surveillance scenarios by adaptively aggregating frame features.

Contribution

It proposes a novel transformer-based aggregation framework that enhances video periocular recognition performance over traditional methods.

Findings

Achieves 99.8% TPR@0.1 false positive rate on COX Face dataset.

Outperforms naive aggregation schemes consistently.

Demonstrates robustness in unconstrained surveillance environments.

Abstract

Video periocular recognition is the task of recognizing an individual's identity based on the region around an individual's eyes. The periocular area is one of the most discriminative regions of the human face, making it suitable for recognition tasks. Its use as a biometric modality has emerged as an alternative, especially in surveillance scenarios where conventional biometric traits such as face or iris recognition become unfeasible due to unconstrained acquisition conditions. This paper proposes an attention-aware approach for video-based periocular recognition in surveillance environments. The framework consists of two main modules: feature embedding and aggregation. The feature embedding module is a deep convolutional neural network that maps periocular data to feature vectors. The aggregation module is an encoder-only transformer that adaptively learns to aggregate frame-level features into a single video representation and a feature vector for the still reference image. Experiments on the publicly available COX Face dataset show the robustness of the proposed method, consistently outperforming naive aggregation schemes. In the best scenario, the approach achieves $99.8\%$ of TPR@$1e^{-1}$ and $96.6\%$ of Rank-5.