When the 1,000-pound gorilla eats your lunch, you eat his breakfast:  Right on the heel of Google’s opening its visual recognition API for free to anyone, Clarifai launches today a free consumers app that can store and classify photo which could very well give a run for its money to Google’s Photo app. Constructed entirely on the company’s award-winning  deep learning visual recognition engine, the  Forevery app might become the favorite companion of every mobile shutterbug out there.

The process is simple. Import your images and the app will automatically tag them

Users can teach the app to recognize customs items, like one's pet
Users can teach the app to recognize customs items, like one’s pet


using the appropriate keyword from the 11,000 + categories it can currently identify. Not just objects such as cars, toy or chair, but locations ( beach, park, mountain) and concepts ( love, happiness, surprise). Your entire library of photos becomes fully searchable, automatically.

Additionally, the app allows you to create custom tags and will learn from them. For example, type in the name of your dog once and the app will automatically tag every other image of your dog. If you named your car, it will do the same. You can teach the system anything you want.

Of course, it has a very efficient and practical face recognition engine for friends and family. All the images of the same person will be clustered together so that you can easily tag one once and the rest will populate by itself. Done.

The Clarifai apps helps you decide what to share with whom based on your past behaviors
The Clarifai apps helps you decide what to share with whom based on your past behaviors

For sharing, computer assistance intelligence comes to the rescue.  The app learns whom you share the most with and what kind of images so over time, it  can accurately suggest what images you should share with whom. No more sharing you drunken party pics with your mom ( not unless you really want to) and no more sharing baby pics with you drinking buddies. As well, recent holiday or kids updates are just a tap away when sharing with your mom.

Finally,  galleries can be created and organized using any of the tags, associating images far deeper than just location and time. For example, around happiness, or smiles. Or shoes and hands. The variations are endless, dynamic and entirely customizable by the user.

Thanks to deep learning visual recognition auto-tagging , Forevery delivers powerful search results
Thanks to deep learning powered auto-tagging , Forevery delivers rich search results

One of the risks, as with any visual recognition solution, is the potential for dramatic false positive that could upset users and create a backlash. Indeed, Forevery makes mistakes and mislabels some of the images,  mostly by being over enthusiasm with labels. Users have the option to edit  the tags, the majority needing just to be deleted. It also has a hard time properly grouping the same people as facial expressions, angles or lighting can play tricks on its facial detection algorithm. Similarly, the interface makes it a breeze to reorganize.   Hopefully, Clarifai has learned from Flickr‘s and Google‘s past experiences and will avoid embarrassing results.

This is the first step into the consumer app world by Clarifai, the visual recognition startup better known for its enterprise API.With 40% of users admitting that they do not have the discipline to sort through their images, the Forevery app will go a long way in helping them stop procrastinating.

Forevery app is free and available on iOS right now.

Photo by InvernoDreaming

Author: Paul Melcher

Paul Melcher is the founder of Kaptur and Managing Director of Melcher System, a consultancy for visual technology firms. He is an entrepreneur, advisor, and consultant with a rich background in visual tech, content licensing, business strategy, and technology with more than 20 years experience in developing world-renowned photo-based companies with already two successful exits.


Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.