Image data is normally used to learn more about the image itself, and often, about its creator. But what if image data could be used to create new images?
Researchers at UC Berkeley have developed software inspired by a technique called image averaging, to reduce thousands of pictures into one, using the “best” qualities of each. According to their paper’s abstract:
“Average images have been gaining popularity as means of artistic expression and data visualization, but the creation of compelling examples is a surprisingly laborious and manual process.”
To fix that problem, they created software to “summarize large amounts of visual data” from sets of images, weighted by “user-indicated performance.”
For example, if Google returned 10,000 image results for the search “cat,” the tool can create one “average” image using the key features of all the photos.
Users can also create subcategories to sort the image results, by giving extra weight to specific features. In this way, they write, “blue-winged butterflies or orange tabby cats might rise to the top of photo collections.”
So, how on Earth could this change shopping (as we suggested in the headline)? Watch, learn—and imagine the future: