Viulib® includes a large number of computer vision techniques and algorithms that together form an efficient and flexible tool to implement advanced solutions in the fields of Computer Vision, Deep Learning and Artificial Intelligence. Viulib® also includes a collection of algorithms and software development techniques that constitute a flexible development platform perfectly suited for tailored-made innovative and disruptive solutions based on specific requirements of the enterprises. Features: • Computer vision based on Deep Learning • Video and image automatic analysis • Real-time operation • Multiplatform, and multi-operating system • C++ implementation • Compatibility with GPU/CPU architectures • Optimized for embedded systems • Modular architecture and simplified dependencies
Deployment and Application
Developments
High speed and dynamic range matrix and line cameras
Line and area sensor based cameras with high spatial resolution and high sensitivity for component surface inspection and dimensional analysis.
Photogrametric image reconstruction system
System for component surface reconstruction based on photometric-stereo technology.
Lane detection and modellization, vehicle detection and classification, traffic sign recognition, pattern detection, driver situation supervision, SLAM (Simultaneous Localization And Mapping)
Anomaly detection in critical infrastructures, real-time face recognition, individual reidentification based on body features, sematic analysis for complex events detection
Non-intrusive vehicle counting and classification, traffic understanding and early detection of accident, people behaviour tracking, multicamera analysis of large infrastructures, 3D inference based on perspective analysis