If you are not great at selfie-taking skills, then researchers from the University of Waterloo have created an app for that. The application does not just enhance the photos you have taken, but provides users with realtime guidance on how to take the selfies. By using the application, users can actually learn how to take better selfies, and understand why they are better. The algorithm directs the user to position the camera to capture the best possible selfie.
Dan Vogel, a professor of computer science at Waterloo says, "Selfie’s have increasingly become a normal way for people to express themselves and their experiences, only not all selfies are created equal. Unlike other apps that enhance a photo after you take it, this system gives direction, meaning the user is actually learning why their photo will be better."
To develop the application, the researchers purchased a database of high resolution 3D scans of average looking people. In a virtual environment, computer generated lighting and a simulated smartphone were used to explore various principles of composition, face position, face size and lighting direction. The researchers then crowdsourced votes for the best selfies, and modelled the patterns of the votes to create an algorithm that suggested the best selfie.
The application was put to the test by getting people to take selfies with a standard smartphone camera app, as well as the smartphone application developed by the researchers. The images were then rated in online surveys. The researchers found that the ratings for the selfies captured with their application was higher than the ratings of photos taken with the standard application. There was a twenty six percent improvement in the ratings for the selfies.
Now the researchers are planning to incorporate more aspects of a selfie into an algorithm, apart from the position of the face and the lighting conditions. The researchers intend to integrate various types of smiles, choice of outfits, and the hairstyle into the algorithm.
Updated Date: Aug 03, 2017 14:14 PM