Visual Sensor Networks

  We are living in a world of camera everywhere and camera on everything. According to the report "iSuppli, Image Sensor Market Tracker, 2011", more than 2,500 million units of CMOS image sensor will be distributed on the market which includes various type of consumer electronics claim to be digital convergence. And we also have plenty of legacy sensors over the world already and many of those are already on mature level to be used in everyday life. 

  We profit from image sensors and legacy sensors as well to make a multi-dimensional context data which will make machine more clever than ever. The visual BigData processing cloud archives and process visual contact data and legacy sensor data as well. By processing multi-dimensional and spatio-temporal fused data, we make machine to understand the visual perception and make camera node intelligence evolving site by site.

   Network capability is now a major up-front feature in state-of-the-art camera systems but mobility of camera node is still out of concern. And the data processing is still done by the hardware bare metal hosts and very stubborn in supporting additional processing algorithms so we cannot expect pleasant maintainability and scalability. Our platform can make a break through with highly scalable architecture and flexible S/W framework. 

  Our visual BigData processing cloud helps to easily scale out the camera aided systems like surveillance camera systems, vehicle blackbox, wearable smart cameras and make them evolvable in diverse processing algorithms like detection and prediction. With the general implementation of cloud infrastructure, we are expecting various applications leveraging visual perception by using cloud intelligence trained by visual BigData.