Serious Computer Vision
Computer vision has never seemed to live up to the hype or potential that Hollywood has expected of it. For many decades there have been images of Minority Report types of use cases that would change the way we all do things. Although, I still hold on to the dream that someday I’ll be able to control my toaster without pressing a button, but I’m not optimistic. At the heart of this inability to live up to expectation isn’t so much that gesture recognition isn’t good enough to enable such use-cases, it’s more that such use-cases don’t really add enough value to the consumer, or solve real-life problems.
In my experience, this mismatch of expectation and execution is typically what happens when a technology is developed without a clear application in mind. 3D sensing technologies such as time-of-flight, or structured light are good at creating 3D depth maps in certain types of environments. However, they’re bulky, expensive and have quite a few environmental limitations.
At EDGE3 Technologies, we don’t think the remote control to the TV should be replaced. It might be fun to show off at my super bowl party how I can wave at my TV, but does that really improve my experience? We are focusing our efforts on what we like to call “serious computer vision problems.” These are problems that improve the safety of automobiles, the efficiency of shipping logistics, or the security of one’s home.
One particular example of this way of thinking is our new mobile box dimensioning product. We have developed an android-based application that leverages off-the-shelf hardware to quickly and automatically determine the dimensions of a box by merely snapping a picture from a mobile delivery device. The reason we built this application is that just this past summer both UPS and FedEx announced plans to begin charging for all ground-shipped packages by size, not just weight, a change designed to increase revenue and reduce costs. Shippers expect this to discourage online retailers from shipping small, lightweight items in large boxes, taking up more space in trucks and raising costs.
Furthermore, as dimensional-weight pricing continues to become the standard in the industry, the requirement will surely find its way to the point-of-sale and effect delivery drivers, retail outlets and consumers. Determining the size of a box, is not only required to determine the shipping price, but also to ensure trucks and airplane are most efficiently loaded.
Surely we can’t rely on the old tape measure to solve this problem?