Movidius’ Vision Processors enable always-on smart capture and continuous vision processing applications for mobile devices.
Mobile imaging, video and vision experiences, highly critical to the end user, are set to be revolutionized in the coming years. This revolution is mainly driven by the needs of software platform owners to introduce massively disruptive application functionality and means of differentiation.
This new paradigm will be a multi-year journey, involving many different implementation approaches and will require a massive amount of parallel computing that still has to fit in the power envelope of a mobile device while driving tremendous innovation in the imaging and vision ecosystem.
The opportunity for Movidius is to provide the best power optimized, yet highly flexible solutions, to enable, together with our partners, a completely new set of vision applications for mobile devices, that up until now were just … impossible to achieve.
Unlike distributed approaches, our architecture maximizes throughput per clock cycle per mW for computational imaging applications thanks to wide data parallelism and an optimized mix of fixed-function hardware and programmable vector processors, all running at modest clock rates in the power sweet spot for the technology process.
Additionally, our memory system architecture allows us to maintain sustained high parallelism with ultra-low latency and lowest power through maximizing data locality.
Movidius, established in 2005, is headquartered in San Mateo, California and operates development centers in Dublin, Ireland and Timisoara, Romania.
Movidius is a venture-backed company with investors including: AIB Seed Capital Fund, Atlantic Bridge, Capital-E, DFJ Esprit and Robert Bosch Venture Capital.