3DiVi has released version 1.15 of its Biometric Anti-Fraud platform, introducing a set of updates focused on privacy compliance, image quality control, and resistance to synthetic media attacks. The company described the release as a major upgrade, with both functional and technical changes outlined in accompanying documentation.
One of the most significant additions in BAF 1.15 is support for per-account biometric data deletion policies. The update allows customers to configure and enforce automated deletion rules for biometric templates at the account level, rather than relying on manual database operations. This capability is positioned as particularly relevant for compliance with data protection regimes such as the EU’s General Data Protection Regulation and US state laws governing biometric data retention.
The release also introduces enhanced image quality assessment features. 3DiVi says the updated quality checks provide more granular feedback on why a facial image fails assessment, including issues related to lighting, blur, and head pose. Customers can configure which quality checks are enforced, allowing them to balance capture strictness against user experience.
BAF 1.15 includes updates to 3DiVi’s web components, incorporating a more advanced face detector and improved head pose estimation. These changes are intended to improve recognition performance when subjects are not facing the camera directly, reflecting real-world capture conditions.
The update further introduces a dedicated deepfake estimation component designed to detect AI-generated or manipulated facial images that could bypass traditional liveness checks. This reflects growing concern that generative AI tools can be used to mount more sophisticated biometric attacks.
Alongside these features, 3DiVi has published detailed release notes documenting API changes, interface updates, and performance optimizations, underscoring the platform’s evolution toward stronger governance and adversarial resilience.
Sources: 3DiVi
—
By the ID Tech Editorial Team
