Yesterday, eBay announced it was snapping up Israeli computer-vision company Corrigon, the ecommerce giant’s fourth acquisition of the year, and 52nd known acquisition of all time.
The deal represents part of a growing trend we’re seeing across the technology realm, whereby machines are used to automate the process of identifying and categorizing images — a task traditionally fulfilled by humans.
For the uninitiated, computer vision is an arm of artificial intelligence (A.I.) that focuses on enabling machines to “understand” images by processing and analyzing them on a pixel-by-pixel basis, rather than relying on human-controlled categorization data, such as keywords and descriptions.
From eBay’s perspective, integrating Corrigon’s smarts into its platform should help it better classify and organize product photos through automation, rather than relying on the accuracy of human input. The deal follows eBay’s recent moves to expand its A.I. capabilities through acquiring startups, evidenced by its acquisition of machine learning company ExpertMaker.
But computer vision, specifically, has emerged as a key focus for companies across the board. Back in March, stock photo giant Shutterstock unveiled a new reverse-image search tool — the first fruits of a computer vision team it had established more than a year previous. The tool uses image recognition to allow anyone to search for pictures based on “look and feel,” (e.g. color schemes, mood, or shapes) rather than using scriptive terms that have been manually applied.
For example, a search for a dog under the old system could surface a whole array of silly alternatives, like this (see similar images at the bottom of the screenshot):
But using computer vision, it’s possible to find visually similar alternatives that may be more suitable.
In the real world
Computer vision has applications far beyond that of image search engines, of course. The burgeoning autonomous driving industry leans heavily on computer vision technology, given that cars must be able to recognize and understand objects in the environment. But the technology is perhaps not quite ready for prime time, as a Tesla S driver was recently killed in a car crash while driving in Autopilot mode. Initial reports suggested that the car may have been unable to recognize the outline of a tractor trailer against the background of its surroundings.
There are many less critical examples that help to illustrate how computer vision is increasingly being used in the real world. Earlier this year, Nanit unveiled a smart baby-monitoring camera that uses computer vision and machine learning to convey insights about a baby’s behavior and sleep patterns.
Elsewhere, Prospera raised a seven-figure sum for its precision agriculture tech that taps computer vision and deep learning to work out how much water to deliver to plants. Prenav, too, raised a notable round for a commercial drone system that helps companies inspect and maintain their physical infrastructure through computer vision.
Big guns from elsewhere in the tech world are also investing heavily in computer vision smarts. Intel recently acquired Movidius to boost its 3D depth camera technology. And, back in August, Facebook announced it was open-sourcing the code for its computer vision algorithms, as it continues to push the boundaries of image recognition technology. “We’re making the code …. open and accessible to all, with the hope that they’ll help rapidly advance the field of machine vision,” explained Piotr Dollar, research scientist at Facebook AI Research (FAIR).
There’s little question that A.I. is one of the hot tech trends to emerge in recent times, and there are many components that will allow it to improve. Brains are great, but eyes give understanding and awareness of the environment and surroundings — real-world, real-time visual data. And that’s why computer vision is integral for artificial intelligence to flourish.