Everpix rethinks how photos are organized

Everpix using semantic understanding to pair similar photos into collections sorted by importance

Everpix, a San Francisco start-up founded just over a year ago, has introduced a new service that uses semantic understanding to highlight and rediscover photos that may be of interest.

If you’re like me, the number of photos that you have taken has multiplied since you began using a smartphone that stores pictures in the cloud. Everpix’s theory is that your photo collection is getting so big that it’s too unwieldy to manage. Even casual users will have more than doubled their collections, it says.

Its solution is a new “highlights” view that scans photos to obtain a sense of its scene or composition, and then it pairs those photos together into collections. It attempts to determine image quality, so that out of focus shots are not displayed, and the best shot is chosen to represent the collection. You can drill down to see the rest.

CEO Pierre-Olivier Latour showed me how Everpix processed images of co-founder Wayne Fan’s recent trip to North Korea and determined which photos were most significant. Those images are displayed larger and more prominently on a tile interface that shows highlights instead of a sequence of every picture that was taken.

Some apps will randomly select photos from a collection or sort by time stamps. I asked for a look under the covers, and was shown a code dump that included things such as GPS data, camera information, and other metadata. It also listed probabilities that an image matched a certain criteria such as a cityscape or natural landscape. That’s the semantic understanding part, which Everpix has spent the past year researching.

Another feature is called Similar Photo Exploration view, which pairs alike photos together regardless of when they are taken. Latour’s iPad Everpix client had grouped urban street scenes that were taken over a period of years. “This is not technology for the sake of technology,” he said. It is intended to help its users resurface forgotten photos.

The technology has the financial support of 500 Startups’s Dave McClure, Picasa co-founder Michael Herf, and Bit.ly investor Mark Hager. Everpix is presently available as native iOS app, as well as a Web service for Android and PC browsers. A Windows file upload service was announced today, along with the official introduction of its highlights view capability. Everpix’s Latou is a former employee of Apple.

Everpix imports photos from Gmail, iPhoto, and cloud storage sites including Flickr. Images are uploaded to its server, hosted in Amazon’s cloud, and viewed through the iOS client software or browser. It costs US$5 per month or $40 for an annual subscription. The iOS app is available for free in the App Store.

(image credit: Everpix)