Reliable detection of doppelgängers based on deep face representations

Doppelgängers (or lookalikes) usually yield an increased probability of false matches in a facial recognition system, as opposed to random face image pairs selected for non-mated comparison trials. In this work, the impact doppelgängers on HDA Doppelgänger and Disguised Faces The Wild databases is assessed using state-of-the-art system. It found that doppelgänger very high similarity scores resulting significant increase match rates. Further, detection method proposed, which distinguishes from mated trials by analysing differences deep representations obtained pairs. proposed system employs machine learning-based classifier, trained with generated utilising morphing techniques. Experimental evaluations conducted Look-Alike Face reveal equal error rate approximately 2.7% task separating authentication attempts doppelgängers.