India’s Unique Identification Authority of India (UIDAI) has published results from its Biometrics SDK Benchmarking Challenge 2025 for fingerprint matching, an evaluation that stands out because it used a long-term child-growth dataset. According to India’s Press Information Bureau, the benchmark tested one-to-one fingerprint verification using samples from children aged 5 to 10 years, with a 5 to 10 year gap between samples, aiming to measure how verification systems hold up as children grow.
Child fingerprints are a known stress case for biometric systems: smaller ridge detail, higher intra-subject variability, and meaningful changes over time can all raise false rejects and push operators toward broader exceptions that weaken assurance. A benchmark built around child-growth data therefore provides a rare primary-source anchor for a question national ID operators often face but rarely quantify publicly: what algorithm performance looks like when enrollment populations include minors, or when continuity over time is a hard requirement.
UIDAI’s summary also points to strong ecosystem interest. The agency said it received 2,106 applications, and after technical screening and security protocols, selected three submissions for final evaluation. It then named Neurotechnology (Lithuania) as the first-place finisher and Innovatrics (Slovakia) as second.
For identity architects, the value of the publication is less about the winners than the operational implications. If a program needs to authenticate minors reliably over time, SDK choice becomes intertwined with lifecycle policy: how frequently biometrics are refreshed, what fallback factors are acceptable, and what exception-handling paths look like when genuine users are rejected. Better verification performance can reduce both access denial and the pressure to reduce controls in the field. UIDAI also said it plans to run similar SDK benchmarking competitions for face and iris modalities, signaling continued third-party testing as a governance mechanism for Aadhaar’s biometric ecosystem.
—
By Ali Nassar-Smith
