Seeing Machines Ltd (AIM:SEE, OTC:SEEMF) lifted the curtain on a major step forward in in-car safety technology, unveiling its next-generation 3D Cabin Perception Mapping platform at CES 2026 last week.
The new system, demonstrated live in Las Vegas, is designed to deliver a real-time digital understanding of everything happening inside a vehicle cabin.
Unlike conventional approaches that rely on separate pipelines for individual features, Seeing Machines’ platform builds a single, high-trust 3D reconstruction of the entire interior, allowing multiple safety and user-experience functions to run from one unified perception layer.
The architecture has been built from scratch and is designed to scale. It supports multiple cameras, multiple occupants and a wide range of features without the need to redesign the system for different vehicle layouts.
By solving for the whole cabin at once, the platform improves consistency and accuracy, even when sensor data is incomplete or noisy.
At the heart of the system is an abstraction layer that separates feature development from camera configuration and raw sensor inputs. In practice, this means software features can be developed once and then deployed across different vehicles and hardware setups.
The company says this cuts development time, reduces cost and speeds up routes to market for new safety capabilities.
John Noble, chief technology officer, described the platform as a fundamental shift in how interior sensing is built and deployed.
He said moving away from feature-by-feature pipelines allows higher accuracy and scalability, while giving customers more freedom to evolve their own safety and user-experience strategies without escalating complexity.
The CES demonstration showcased a live 3D reconstruction of a full vehicle cabin using three cameras covering three rows of seats and up to seven occupants.
The system tracked body size, shape and full 3D pose for every occupant, including height and weight classification. It also demonstrated detection of unsafe seating positions, child seats, seat configuration and random objects such as bags and phones.
While the initial focus is automotive, Seeing Machines said the technology has wider potential. The same perception layer could be applied to robotics and other human-machine interaction environments where accurate understanding of people and space is critical.
The architecture also allows different 3D sensing technologies to be mixed and matched as hardware evolves.