COMPUTER VISION 2017: THE YEAR IN REVIEW

Computer Vision

Noteworthy news

Some Meats Still a Mystery for Food Recognition APIs

Some Meats Still a Mystery for Food Recognition APIs

Grubhub Bytes (Medium)

Image recognition–particularly of the API variety–is still a work in progress, a state of affairs revealed by food delivery app Grubhub’s evaluation of three popular computer vision APIs. Google Vision API, Amazon Rekognition and Microsoft Computer Vision API all have their strengths and weaknesses, but none of them seem to be able to identify a samosa, chicken wings, or fruit salad by name, as any human might. It sounds like these APIs need to go on a diet–of more food porn.   

Computer Vision to Elevate Your Instagram Feed

Computer Vision to Elevate Your Instagram Feed

Digital Trends

If you’ve been gazing enviously at those super popular Instagrammers with aesthetically consistent feeds, there’s an app for you. Mosaic uses image recognition and other assorted AI to scan and organize your images around recognizable themes like “vintage” or “light,” then lets you update them on your feed. New themes will be added regularly, giving big-time influencers plenty of time to stay ahead of the machines, themes-wise.

Are Computer Vision Smartphone Cameras Ready for Their Closeup?

Are Computer Vision Smartphone Cameras Ready for Their Closeup?

The Verge

It sounds like the new de rigueur feature for smartphones is a computer vision-enabled camera that identifies what’s in the picture you’re shooting–say food or a flower–and then automatically optimizes light and shutter settings accordingly. LG is set to announce a built-in AI feature for its upcoming V30 smartphone that’ll not only improve your pictures, but also include visual search for live objects and shopping. No word yet on whether this feature will be cloud-based, requiring a phone signal, or local on the phone, as on the just-released Huawei Mate 10 Pro, which uses an onboard chip to pull off computer vision-enabled camera optimization even in airplane mode.

Silly Car-Into-Mouse Conversion Is Also Strangely Smart

Silly Car-Into-Mouse Conversion Is Also Strangely Smart

William Osman and Simone Giertz (YouTube)

As William Osman, the mechanical engineer-cum-maker who made this hilarious video with “crappy robot queen” (and fellow YouTube star) Simone Giertz, points out, the optical mouse is the “cheapest optical flow sensor you can buy.” This thanks to an LED light that bounces off the surface and is registered by the mouse’s built-in camera and analyzed by a digital signal processor to figure out the movement of the mouse (then sends that info to the computer so you can control your desktop). It’s a rudimentary form of computer vision that was cleverly modified with a new lens and a magnetometer to transform a DIY electric car into a computer mouse that must be driven around a parking lot to send a simple email. Silly sounding, for sure, but somehow clever as well, in a cool-new-uses-for-old-things meets real-yet -virtual-cockpit kind of way. Practicality aside, hilarity ensues in the tinkering pair’s videos.

View more news >>

The AI-Human Collaboration at Work: Q&A With Albert Technologies CTO Tomer Naveh

Artificial Intelligence

Stories

Want to Learn About AI? Follow These 5 Robotic YouTube Stars

Want to Learn About AI? Follow These 5 Robotic YouTube Stars

Artificial Intelligence

The Interactive Guide to Teaching Machines to See the Way We Do

The Interactive Guide to Teaching Machines to See the Way We Do

Advertising

Eight Times Computer Vision Hilariously Failed

Eight Times Computer Vision Hilariously Failed

#FAIL

Want to Know How AI Is Transforming the Media Industry? Just Ask Our Chatbot

Want to Know How AI Is Transforming the Media Industry? Just Ask Our Chatbot

Artificial Intelligence

5 Ways Computer Vision Is Giving the Fashion Industry a Makeover

5 Ways Computer Vision Is Giving the Fashion Industry a Makeover

Computer Vision

Moneyball 3.0: How Visual Data Is Supercharging Sabermetrics in Sports

Moneyball 3.0: How Visual Data Is Supercharging Sabermetrics in Sports

Computer Vision

View more stories

Guides

Read more