This study evaluates 17 Face Image Quality Assessment (FIQA) methods as signals for detecting morphing attacks. On the FRLL dataset, DifFIQA(R) achieved a mean D-EER of 11.42%, demonstrating that quality scores can identify manipulated images without morph-specific training.
HOW THIS AFFECTS YOU
●
builderYou can integrate existing FIQA models into security pipelines to detect manipulated facial data.
●
policyThis offers a training-free, scalable method for enhancing biometric security against sophisticated morphing attacks.