14 DAY TRIAL // Have you ever wondered what the world around you would look like if you drew swaths of pink ink across your house or how your car would change color with a single finger swipe?
Wonder no more as Microsoft has developed an impressive tech demo called "SemanticPaint" that lets you scan objects in 3D using Kinect. Even though this is not anything new, it does, however, let you separate and define individual objects nearby you.
Using this technology AIs and self-driving cars will be able to differentiate between all sorts of objects and define individual objects in the scene. To demonstrate this even further the researchers scanned the context with a Kinect, then have it introduced in the computer as a single object, afterwards a person will come and define each object separating them from that singular entity in real time, by naming them. The computer will then separate them in classes. As processing power for this software an average laptop CPU is used for better interactivity.
However, more powerful systems are used in the background to help the software label and deduce if one object belongs to a specific class like "table" or "chair." Microsoft claims that the software has a learning ability, remembering and differentiating between labels in order to perform tasks better and faster in the future.
Microsoft believes this system will prove crucial in developing robot guidance software and creating detailed maps of a scene complete with individual objects. But probably the main purpose of this software is to actually help blind people see through ophthalmologically augmented realities.
