Development // Consulting // Research

15 years of experience in Computer Vision technology

For seamless integration of virtual objects into video, highly accurate camera parameters are needed. We provide software for accurate camera tracking and calibration.

For training AI, our vision based tools build datasets including image segmentations, bounding boxes, or 3D shapes with minimal manual interaction. Using uncertainty measures, we care for data quality and dataset balancing.

We provide various methods for object detection and tracking. In the 5GCAR EU research project, our software tracks vehicles from road-side cameras in real-time. For special object types, training data is constructed.

For the planning of refurbishment, Building Information Modeling (BIM) is required. In the DIGI-PV research project, building façades are reconstructed from camera images with a high degree of automation. Images are captured with a hand-held camera from the ground or with a drone. Our approach reduces time & cost to build Digital Twins.

We fuse classical Computer Vision technology with AI capabilities for object detection and segmentation. Use the best of both worlds!