Over the past year, one of our engineering teams has spearheaded and modified computer vision technology to introduce more innovative search and discovery features and to improve your overall site experience.
In the past, Shutterstock’s search algorithm and similar image offerings on image pages were powered by keywords provided by our contributors when they upload their content. That keyword data, while useful for indexing images into categories on our site, wasn’t nearly as effective for surfacing the best and most relevant content. So our computer vision team worked to apply machine learning techniques to reimagine and rebuild that process.
The technology now relies instead on pixel data within images. It has studied our 70 million images and 4 million video clips, broken them down into their principal features, and now recognizes what’s inside each and every image, including shapes, colors, and the smallest of details; this visual and conceptual data is represented numerically.
Why does that matter? All the data Shutterstock has collected about our content has led to the creation of an unparalleled reverse image search. You can now upload an image of your choosing to Shutterstock’s site, bypassing the need to type a query into a search bar, and the technology will identify similar images in look and feel inside of Shutterstock’s collection.
Shutterstock will also soon launch visually similar discovery for video. We’re just getting started with what this technology can do.
With reverse image search you can now use a photo or illustration to find other images with a similar look and feel. Simply drag an image into the search bar and you’ll get results based on pixel data instead of the standard keyword data. See a demo of reverse image search below.
Shutterstock has also improved visually similar results for images and footage. In this example, the old similar images are sourced from keywords, giving you different compositions, styles, and subjects.
The new visually similar tool is based on pixel data, giving you highly related results that match in color, subject, and composition.