Google Clips camera uses AI to play photographer
With the days of needing to store photo and video on film well and truly behind us, sometimes it can feel like the biggest challenge with a camera in hand is sorting the wheat from all the chaff. A new Google camera called Clips aims to avoid this dilemma by removing it from our hands entirely, leaving it up to AI to determine the moments that need capturing in the first place.
Modern smartphones have perfectly adequate, if not excellent cameras, and enough storage space to keep even the biggest selfie-enthusiasts snapping away. So why would anyone need second, similarly capable camera? Because artificial intelligence, that's why. Google thinks its machine learning software can recognize those times in our life that need capturing, leaving us free to enjoy the moment and the photographic evidence thereafter.
This type of technology has been in the pipeline for a few years now. Last year we looked at household camera called the Kiba that was triggered by visual and audio cues, and then used something called a Joy Ranking Algorithm to select the highlights, edit them into a home video, and discard everything else.
The QuikStories feature on GoPro's mobile app works in a similar way, automatically skimming footage and editing the highlights into a ready-to-watch video, even matching the transitions to music. Other action cams like the Graava promise to do this on the fly, using built-in sensors to recognize notable events, like a sudden change in speed or orientation. Nest's security cameras do much the same thing in an effort to save homeowners scanning hours of footage.
With Clips, Google is trying to steer this toward everyday applications where spontaneous, capture-worthy moments can pop up and you don't have the time or desire to whip out your smartphone. It can be set down on a flat surface or clipped onto something, like the back of a chair, and once switched on, remains on the lookout for moments to capture, be it a recognizable face or a stable, clear view of the scene.
These are recorded as short motion photos just a few seconds long and are synched with a paired Android or iOS smartphone. There they can be edited, saved and deleted, or individual frames can be pulled as high-res photos. Google says the camera gets better over time at recognizing faces, with all the machine learning taking place on the camera itself.
Clips appears more a playful way of testing the waters with always-on, automated cameras than simply allowing the entire family to be in frame (hello tripod, hello selfie stick). And that mightn't sit well those preoccupied with their privacy. For what its worth, Google does emphasize that Clips is intended for use in the home, and that it lights up when on so everyone knows they are a chance of being recorded.
It hasn't specified a release date for Clips, but does say that it is coming soon to the US and will be priced at US$249.