Scene Extraction from real world data

Flexa, with over 15 years of expertise in AI and Cloud, delivers this solution to create measurable business impact.

Our platform runs directly in the customer’s AWS account, ensuring scalability, compliance, and full ownership.

During the development of ADAS/AD (Automated Driving) features, test drivers collect data in various locations around the world. Typically, this data can only be used for projects that share a similar sensor setup. If the sensor configuration differs significantly — for example, if the cameras operate at a higher resolution or additional radars are integrated into the system — new data must be collected.

Generative AI can be used to extract detailed scene information and convert it into a text-based format such as ASAM OpenSCENARIO. This structured data can then serve as input for simulations in virtual environment tools such as IPG’s CarMaker, Hexagon’s VTD, or Vector’s DYNA4.

Designed to reduce costs, accelerate development processes, and enhance both customer and employee experiences.

Scalable and secure, compliant with enterprise-grade standards.

Enables strategic implementation with a focus on continuous improvement.

The Flexa solution is powered by Amazon Bedrock and Amazon SageMaker, enabling robust, scalable, and intelligent automation tailored to this use case.