homefocus areas › video and image analytics


video and image analytics   back to focus areas

Bringing the sense of sight into the fold of data analytics

Human vision is a rich and powerful sense that lets us recognize and interpret vast amounts of information with just a glance. It is so completely integrated into our brain and thoughts we rarely pause to notice how amazing it is.  If you had to rely on hearing, tasting, touch and smell - along with physics and algebra - to drive a car, your life would be spent in the driveway. Add vision, and instantly, without effort, your brain recognizes the constantly changing scene before your eyes and directs the body to brake quickly to avoid hitting the truck in front of you.  It is a marvel.  

Cameras create physical snapshots of what we see. Over the centuries these have evolved from crude, static monochrome records to high-resolution digital video. But the interpretation of that video is still largely a chore for the brain. Today’s vast computer infrastructure still struggles to provide meaningful analysis of video in real time.

Video and image analytics rapidly are eliminating these limitations, opening up opportunities to bring the benefits this vision to our daily lives.

With vision, machines will move beyond simply carrying out specific, well-defined tasks. They will be able to "see" and extract information from images and video on their own, dynamically and continuously learning from, and accumulating knowledge in much the way humans do.

Innovation needed at two levels

 Fundamental scientific advances are needed to add the sense of sight to data analytics in a way that will create novel and high-value applications.

PARC research is at the cutting edge in many of these capabilities including detection, recognition, and tracking of people and objects from multi-camera systems; analysis of human behavior; understanding and acting upon video from a variety of systems, such as mobile and wearable platforms and fast identification and access of rare or unexpected events.

 Innovation also is needed to transform core scientific capabilities into practical, reliable, and effective solutions to the real-world problems faced by our customers. This type of systemic, client-centric innovation is a core strength at PARC, with recent successes including license plate recognition and vehicle occupancy detection, systems that are already deployed and generating revenue for clients.

Imagine isolating valuable information from hundreds of hours of video without human intervention. At PARC we believe potential applications for video analytics span a broad range of industries. To that end we are working with clients across a number of different domains to bring the sense of sight into the fold of data analytics.