发表机构
University of Nebraska Lincoln（内布拉斯加大学林肯分校）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
该研究针对低质量人脸识别的挑战，提出结合局部概率间隔、嵌套注意力模块和质量门控协议的统一框架，在多类基准上实现了识别与验证性能的一致提升。
AI 中文摘要
低质量人脸识别（LQFR）仍然是一个挑战，因为难以将退化的查询（probe）图像与低质量（LQ）注册（gallery）图像匹配，且大规模模型的训练数据稀缺。尽管最近的人脸识别（FR）模型在高质量（HQ）图像上表现良好，但在信噪比（SNR）极低的LQ图像上，其准确率会显著下降。此外，在LQ数据上对HQ预训练模型进行微调通常会提高LQ识别性能，但会以牺牲HQ泛化能力为代价，这种权衡在跨越多个不同图像质量水平数据集的现代评估场景中更为明显。为解决这些限制，我们提出了一个统一框架，包含三个主要组件：（1）局部概率间隔（LPM），直接从模型的判别空间估计每个样本的难度；（2）嵌套注意力模块（NAM），一种新的低秩适配器模块，在选定的Transformer层中嵌入自注意力机制；（3）质量门控协议（QGP），利用现成的图像质量估计器在测试时调节适配器的贡献，使单个模型能够处理全质量范围而不牺牲HQ性能。在监控数据集（TinyFace、SurvFace）和标准人脸识别基准（IJB-B、IJB-C）上的实验表明，该模型在识别和验证任务中均取得了一致的性能提升。代码和模型将在此http URL发布。
英文摘要
Low-quality face recognition (LQFR) remains challenging due to the difficulty of matching degraded query (probe) images against low-quality (LQ) enrollment (gallery) imagery and the scarcity of training data for large-scale models. While recent face recognition (FR) models perform well on high-quality (HQ) imagery, their accuracy drops significantly on LQ images with extremely low signal-to-noise ratio (SNR). Moreover, fine-tuning HQ-pretrained models on LQ data often improves LQ recognition at the expense of HQ generalization. This trade-off becomes more pronounced in modern evaluation settings spanning multiple datasets with varying image quality levels. To address these limitations, we propose a unified framework that combines three main components: (1) Local Probability Margin (LPM), which estimates per-sample difficulty directly from the model's discriminative landscape; (2) Nested Attention Module (NAM), a new low-rank adapter module that embeds a self-attention mechanism within selected transformer layers; and (3) Quality Gating Protocol (QGP), where an off-the-shelf image quality estimator modulates the adapter contribution at test time, enabling a single model to handle the full quality spectrum without sacrificing HQ performance. Experiments on surveillance (TinyFace, SurvFace) and standard (IJB-B, IJB-C) face recognition benchmarks demonstrate consistent gains in both identification and verification. Code and models will be released at github.com/candllq/nam.
CommentsAccept at IEEE/IAPR IJCB 2026
