Superior vision: Always on human-like vision for intelligent devices

摘要

We live in a world of intelligent robots, drones and mobile phones/assistants. The technologies behind these products are heralded by a new wave of deep learning applications ported on mobile devices. The robotic and vision technologies behind these products will shift the applications of electronic devices to a more superior level of intelligence that will change our world. From a hardware point of view, these devices will become a hub of learning sensors that will be able to capture and learn from their surrounding environment and their context. On top of the aforementioned deep learning technologies, the most important sense for these devices will be their vision capabilities. To make this sense as close as possible to human vision, will require advanced technology to capture, process, understand and provide analytics in real time to enable their applications make critical ???educated: decisions???. ???Always on??? vision capabilities are envisaged as the paradigm for the next generation of mobile This Keynote will cover in this context the state of the art in the domain of vision with the latest image processing, computer vision and deep learning algorithms implemented on various low-power processing architectures.