When people go to a hair stylist, they often browse magazines to find a photo of a style they want to try, but they won’t know if they really like it until they try it.
Police are often challenged when looking for missing people when those individuals disguise their appearance with different hair colors and styles or facial hair. And children who have been missing for years can be very difficult to find because their looks have changed over the years.
What if we had help figuring out all these appearance challenges?
A new system called Dreambit, developed by a University of Washington computer vision researcher, lets a person imagine how they would look with a different hairstyle or color or even in a different time period, country, or anything that can be queried in an image search engine.
After uploading an input photo, you type in a search term (such as curly hair, India, 1930s), and the software's algorithms mine Internet photo collections for similar images in that category and seamlessly map the person's face onto the results.
Dreambit draws on previous research conducted at the UW and elsewhere in facial processing, recognition, three-dimensional reconstruction, and age progression, combining those algorithms in a unique way to create the blended images.