Tracking and recognition
We work with visual tracking and recognition in uncontrolled environments. We focus on excellent performance even in crowded situations and difficult weather conditions, dealing with complications like shadows, specular reflections and motion blur. This technology is the basis for applications like pedestrian detection for active automotive safety, intelligent surveillance systems and safety systems for critical infrastructure. Our platform contains some of the best-performing techniques for visual tracking, video stabilization, shadow suppression and object recognition available.
We actively follow current research in deep learning. This technology has provided ground-breaking performance gains for challenging vision tasks like large-scale object recognition and whole-scene analysis. We have experience in applying open-source libraries for deep learning in practical applications, and know how to integrate such tools in client products. Our platform contains reusable code blocks for integration and adaptation of deep learning systems.
We have developed tools used in cancer diagnostics from MRI images, and have experience in most technologies popular in medical imaging. Our platform include some of our condensed experience within image registration, segmentation and contour analysis, in both 2D and 3D.
Image and video quality
We have experience from working with noise reduction and image quality for challenging air-traffic control applications and with image sensor calibration for high-end professional digital cameras. We also have experience in video compression, high dynamic range imaging and custom-made automatic exposure control.
We developed core image analysis code for one of the world’s largest vision sensor manufacturers. We also worked with machine builders, developing custom solutions for specific visual inspection applications.