High-Fidelity Synthetic Data for In-Cabin Monitoring AI

Anyverse works in close coordination with Euro NCAP to implement high-fidelity synthetic data for the assessment of in-cabin monitoring systems. Our synthetic datasets include parameterized scenarios, biomechanical human models, and physics-based sensor simulation (RGB, IR, radar), enabling reproducible, standardized evaluation across DMS, OMS, and CPD systems. This collaboration aims to define clear performance criteria and streamline validation protocols, helping OEMs and Suppliers design and validate safer in-cabin AI ahead of the 2029 standard.

As part of this effort, Anyverse has launched the Virtual Assessment Implementation Program (VAIP) — a technical program operationalizing the Euro NCAP virtual assessment framework for in-cabin monitoring systems. Learn more about the VAIP here.

How does the Euro NCAP collaboration benefit OEMs and suppliers?

The collaboration helps OEMs and Tier 1 Suppliers adopt validation processes aligned with evolving Euro NCAP requirements. By using Anyverse synthetic simulation, teams can test DMS, OMS, and CPD systems against standardized, repeatable scenarios, reducing uncertainty, compressing testing cycles, and accelerating compliance readiness for both current 2026+ development cycles and the 2029 deadline.

OEMs and Suppliers looking for a structured path to implement the Euro NCAP virtual assessment framework can apply to join the VAIP.

Anyverse INCABIN is a web app that generates synthetic data for driver monitoring use cases. It is powered by Anyverse’s high-fidelity synthetic data generation platform for computer vision. Unlike general-purpose 3D engines or simulation tools, Anyverse InCabin provides access to all test cases evaluated by Euro NCAP, enabling users to generate targeted data that supports system validation and helps achieve high safety ratings. The platform focuses specifically on sensor-accurate, scenario-based data generation for Driver Monitoring Systems (DMS), Occupant Monitoring Systems (OMS), and Child Presence Detection (CPD). It supports Radar, RGB-IR, NIR, and LiDAR outputs with pixel-perfect annotations, making it ideal for AI training and pre-validation before final validation in real-world environments.

Driver Monitoring AI systems often need to handle rare, privacy-sensitive, and safety-critical scenarios like drowsy driving, distracted behavior, or unattended children. Synthetic data allows teams to simulate these situations at scale, with perfect labels and full environmental control. Anyverse lets you create this data quickly, safely, and with complete diversity, removing the need for costly and ethically complex real-world data collection.

Anyverse offers a dedicated library of synthetic test cases aligned with Euro NCAP protocols, including those for Driver Monitoring Systems (DMS), Occupant Monitoring Systems (OMS), and Child Presence Detection (CPD). OEMs and Tier 1 suppliers can generate datasets that directly support the training, testing, and validation of in-cabin AI systems required to meet or exceed Euro NCAP safety benchmarks.

Anyverse covers a broad set of test cases relevant to Euro NCAP’s roadmap, including:

• Driver drowsiness and inattention detection
• Distraction due to mobile phone usage or object interaction
• Driver and front-seat passenger seatbelt usage
• Child left unattended in the rear seat
• Occupant presence and classification per seat
• Behavioral patterns indicating medical emergencies

These scenarios are fully customizable and can be reproduced across varying lighting, sensor types, and occupant profiles.

Anyverse supports multi-sensor simulation, including RGB cameras, near-infrared (NIR), Radar, and LiDAR. This ensures realistic outputs that mirror the performance of production-grade sensor stacks in modern vehicles.

Yes. Anyverse includes a scenario-driven approach to data generation. You can define and simulate behaviors like driver yawning, using a mobile phone, unbuckled seatbelts, children moving in rear seats, or passengers partially occluded by objects. This allows AI models to train on real-world complexity with synthetic data accuracy.

Anyverse offers extensive demographic and behavioral diversity. You can generate synthetic humans of different ages, ethnicities, genders, body types, and clothing/accessories. Expressions, gaze direction, posture, and movement can all be controlled to reflect real-life variability. Environments can be modified for different lighting, camera angles, and car interiors.

Anyverse delivers automated, pixel-perfect annotations, including:

• 2D and 3D bounding boxes
  
• Semantic and instance segmentation masks
  
• Facial landmarks and head pose
  
• Keypoints for body and hand tracking
  
• Object presence (e.g., phone, cup, bag)
• Occupancy detection per seat

Our proprietary and comprehensive annotation output can be easily transformed into COCO, KITTI, and custom formats compatible with leading computer vision frameworks.

Anyverse enables scenario-based data generation, which fits perfectly into data-centric AI workflows. As you train models and identify weaknesses (e.g., poor night performance or trouble detecting sunglasses), you can generate new synthetic data that targets those gaps. This reduces dataset waste and shortens training cycles.

Anyverse is trusted by OEMs and Tier 1 suppliers working on autonomous driving, driver monitoring systems (DMS), and occupant safety technologies. While many clients are confidential, the platform has been successfully used to develop production-grade computer vision models for automotive environments.

Anyverse InCabin is designed to be accessible to everyone, no coding or advanced technical knowledge required. Users can simply select their desired DMS, OMS, or CPD use case, follow the configuration steps, and generate a dataset in minutes. For more tailored needs, we offer full customization, including modeling any car cabin or interior, support with scenario creation, sensor tuning, and expert guidance for dataset design. Our team can help replicate your exact vehicle interior and sensor setup to ensure simulation realism and domain relevance. We also provide onboarding, continuous support, and dataset iteration services to accelerate your development process.

Getting started is easy. You can request a demo tailored to your use case through our website, whether you’re building a Driver Monitoring System (DMS), implementing child presence detection (CPD), or Occupant Monitoring Systems (OMS). We’ll show you how Anyverse can dramatically reduce data acquisition time while improving model robustness.

After we’ve analyzed your data needs—and before you commit to an Anyverse InCabin subscription—you can also request a free trial*.

*The Anyverse InCabin trial version allows users to access the app for a limited period of time. Please note that the trial is not the full version; some test cases and features may not be available.