Humans have long been masters of dexterity, a skill that can largely be credited to the help of our eyes. Robots, meanwhile, are still catching up. Certainly there's been some progress: for decades robots in controlled environments like assembly lines have been able to pick up the same object over and over again.
Recent breakthroughs in computer vision have allowed robots to make basic distinctions between objects, but even so, robots cannot really understand the shapes of objects, so they can't do much to pick them up quickly.
MIT computer science and artificial intelligence laboratory (CSAIL), the researchers reported in a new paper, they have made great progress in the field, the system can let robot to check random objects, and visually know enough about them, and in the case of had never seen them perform specific tasks.
The system, dubbed "Dense Object Nets" (DON), looks at objects as collections of points that serve as "visual roadmaps" of sorts. This approach lets robots better understand and manipulate items, and, most importantly, allows them to even pick up a specific object among a clutter of similar objects -- a valuable skill for the kinds of machines that companies like Amazon and Walmart use in their warehouses.
For example, someone might use DON to let a robot grasp a specific position on an object -- like a tongue.Since then, it could see a shoe it had never seen before and managed to grab its tongue.
The team looks at potential applications not only in the production environment but also in the home.Imagine giving the system an image of a tidy house to keep it clean while you work, or using an image of a plate so the system can put away your plate while you're on vacation.
What's also noteworthy is that none of the data was actually labeled by humans; rather, the system is "self-supervised," so it doesn't require any human annotations.