The New York-based company offers more than 80 million images for bloggers and media outlets, but keyword searches aren’t always the most effective way to find images relevant to a story.
If you want to search for photos that are similar to ones you already have in your possession, or if you want to find alternative photos based on the shapes, mood, color scheme, and general mise en scène around you, reverse image search comes into play. You can search Shutterstock by using the camera on your iPhone or the photos on your camera roll to find similar images.
The launch comes three months after Shutterstock first introduced the feature through its desktop version, though extending it to smartphones does feel like a natural move, given that smartphones are cameras in their own right.
“When we unveiled Reverse Image Search this past spring, we knew that it was a perfect fit for our mobile application — it’s arguably one of the best use cases for computer vision technology, in general,” said Shutterstock CEO and founder Jon Oringer.
Computer vision is an arm of artificial intelligence that lets machines search for images by processing and analyzing them on a per-pixel basis, rather than through meta data, such as keywords and descriptions that rely on manual human input. The reverse image search feature was the first output from Shutterstock’s computer vision team, which was put together a little more than a year ago.
To use the new feature within the Shutterstock iOS app, tap the search box, and you’ll see options to search by snapping a photo then and there, uploading one from your camera roll, or even using cloud-based services such as Dropbox and iCloud. Once you’ve selected a photo, Shutterstock takes a few seconds to process the request and then presents the results in a grid-like format.
Reverse image search is nothing new, of course — Google has offered the feature for some time and recently expanded its service to cover mobile searches. But as a service aimed specifically at photographers, bloggers, and journalists, Shutterstock’s latest smarts could prove particularly useful.
As with all machine-learning technology, it should improve over time. It should also allow Shutterstock tap some useful data. “As users upload photos captured by phones to search Shutterstock’s collection, the neural network on Shutterstock’s back-end studies and learns what types of images are most popular for mobile usage rather than desktop usage,” the company said. “With time, it will grow to understand authentic photography taken in more natural settings. Data collected will showcase emerging trends and best techniques on mobile devices.”