Face Verification with Challenging Imposters and Diversified Demographics
Adrian Popescu (1), Liviu-Daniel \c{S}tefan (2), J\'er\^ome, Deshayes-Chossart (1), Bogdan Ionescu (2) ((1) Universit\'e Paris-Saclay,, CEA, List, Palaiseau, France, (2) University Politehnica of Bucharest,, Romania)

TL;DR
This paper introduces the FaVCI2D dataset to address limitations in existing face verification datasets by including challenging imposter pairs and diverse demographics, revealing performance gaps in state-of-the-art models.
Contribution
The creation of the FaVCI2D dataset with challenging imposter pairs and demographic metadata, highlighting the need for more robust evaluation in face verification.
Findings
State-of-the-art models perform significantly worse on FaVCI2D.
Existing datasets are too simplistic and lack demographic diversity.
Performance drops confirm the need for more challenging benchmarks.
Abstract
Face verification aims to distinguish between genuine and imposter pairs of faces, which include the same or different identities, respectively. The performance reported in recent years gives the impression that the task is practically solved. Here, we revisit the problem and argue that existing evaluation datasets were built using two oversimplifying design choices. First, the usual identity selection to form imposter pairs is not challenging enough because, in practice, verification is needed to detect challenging imposters. Second, the underlying demographics of existing datasets are often insufficient to account for the wide diversity of facial characteristics of people from across the world. To mitigate these limitations, we introduce the dataset. Imposter pairs are challenging because they include visually similar faces selected from a large pool of demographically…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Code & Models
Videos
Face Verification with Challenging Imposters and Diversified Demographics· youtube
Taxonomy
TopicsFace recognition and analysis · Face Recognition and Perception · Evolutionary Psychology and Human Behavior
