Book a demo
Research & KnowledgeEvent: September 8, 2026

AI Verse provided data for a FOV drone detector model achieving 0.989 mAP50 with no frames captured.

Published

Signal category

Research & Knowledge

Quote

AI Verse data provided data for FOV drone detector model of 0.989 mAP50. Not one frame was photographed.

AI Verse team

Company

AI Verse

The synthetic data procedural engine powering defense and security AI.

Industry
Software Development
Location
BIOT, FR
Company size
22 employees

AI Verse is a French deeptech company (launched in INRIA, a NATO DIANA company) that generates annotated synthetic imagery to train computer vision AI models. This improves model robustness and coverage in scenarios where real-world data is scarce, expensive, or incomplete. The platform operates on sovereign infrastructure, available as SaaS or deployed on-premise, with no dependence on foreign cloud services, meeting classified data security requirements. Key use cases: * Detection, recognition, classification, and tracking across EO and satellite imagery (geointelligence) or infrared sensors (LWIR/MWIR/SWIR). * Applicable across land, air, maritime, and space domains, including: * Counter-UAS operations (detection and interception) * Threat detection (vehicles, personnel, weapons) * Inter-drone perception and swarm detection * Vision-based geolocation and navigation in GPS-denied environments * AI Verse data supports two approaches: * Standalone training on purely synthetic data * Or augmenting existing datasets, generating rare targets, degraded conditions, sensor-specific variations, etc. Beyond image generation, the platform is expanding this year into video and trajectory simulation, for instance, to train and validate autonomous drone systems (path planning, threat detection, target tracking).

Founded 2020

Customize signals for your business.

Know everything happening across the B2B world, and act on the company movements that matter to you.

© 2026 SeedOpsCompany intelligence.

SeedOps.

AI Verse provided data for a FOV drone detector model achieving 0.989 mAP50 with no frames captured. | SeedOps