Vision-language models enable zero-shot face image quality assessment whose biometric utility depends on model architecture rather than size, with outputs that align with traditional methods but vary by prompt.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Employing Vision-Language Models for Face Image Quality Assessment
Vision-language models enable zero-shot face image quality assessment whose biometric utility depends on model architecture rather than size, with outputs that align with traditional methods but vary by prompt.