ATTN-FIQA: Interpretable Attention-based Face Image Quality Assessment with Vision Transformers

TL;DR

ATTN-FIQA introduces a training-free, attention-based face image quality assessment method leveraging pre-trained Vision Transformers, providing interpretable quality scores without additional training or modifications.

Contribution

It demonstrates that pre-softmax attention scores from Vision Transformers can serve as effective, interpretable face image quality indicators without extra training.

Findings

Attention scores correlate with face image quality across benchmarks.

The method requires only a single forward pass, no retraining.

Attention patterns reveal facial regions influencing quality assessment.

Abstract

Face Image Quality Assessment (FIQA) aims to assess the recognition utility of face samples and is essential for reliable face recognition (FR) systems. Existing approaches require computationally expensive procedures such as multiple forward passes, backpropagation, or additional training, and only recent work has focused on the use of Vision Transformers. Recent studies highlighted that these architectures inherently function as saliency learners with attention patterns naturally encoding spatial importance. This work proposes ATTN-FIQA, a novel training-free approach that investigates whether pre-softmax attention scores from pre-trained Vision Transformer-based face recognition models can serve as quality indicators. We hypothesize that attention magnitudes intrinsically encode quality: high-quality images with discriminative facial features enable strong query-key alignments producing focused, high-magnitude attention patterns, while degraded images generate diffuse, low-magnitude patterns. ATTN-FIQA extracts pre-softmax attention matrices from the final transformer block, aggregate multi-head attention information across all patches, and compute image-level quality scores through simple averaging, requiring only a single forward pass through pre-trained models without architectural modifications, backpropagation, or additional training. Through comprehensive evaluation across eight benchmark datasets and four FR models, this work demonstrates that attention-based quality scores effectively correlate with face image quality and provide spatial interpretability, revealing which facial regions contribute most to quality determination.