Subscriber Discussion

Crowdsourcing Object Recognition - Hope Or Hype?

In the aftermath of the tragic disappearance of Malaysia Flight MH370 bound for China, Tomnod has established a crowdsourcing effort to detect the downed plane or its wreckage, by volunteers poring over satellite imagery of the most likely incident areas.

My question is: Is the state of our finest world class image analytics such that we have a greater chance of success using untrained humans than our fastest computers with their best algorithms? Isn't it easier to detect a submerged plane, or strewn flotsam/jetsam than whether a person is acting suspiciously in a crowd?

Do you think they are using any analytics from any of the companies were used to hearing about, i.e. Video IQ, Aimetis, 3VR, BRS Labs etc?

If BRS had been on the ball, they would have detected the wreckage before the plane even took off...

There's a whole different class of technology / companies that specialize in satellite /aerial image analysis. I recall OV doing some projects for the US government in this area but typically video analytic companies in physical security do not cross into this area.