We all at least once have had an otherwise great photograph ruined because the lens snapped exactly at the time you blinked. To address this ordeal, Facebook has launched a system that replaces closed eyes with open ones in a remarkably convincing manner, using the capabilities of artificial intelligence (AI).
In-painting is a technique where a program fills in space with what it thinks belongs there. Adobe, in particular, has made good use of this technique with its "context-aware fill" which allows users to replace undesired features. However, some features are beyond the tool's functions, which also include the closed eye!
Facebook, which has more pictures uploaded on its platform than any other in the world, uses a machine learning system called Generative Adversarial Network (GAN).
In GAN, one part of the system learns to recognize and another part of the system repeatedly creates images that progressively grow realistic based on the feedback from the recognition part, say, face for example.
According to a report in TechCrunch, existing systems left something undesired. They appear to paste in the eyes of the people without much consideration for compatibility with the rest of the image. For that matter, they have no intuitive perception of eyes, skin colour or anything at all.
However, Facebook’s researchers include "exemplary" data showing the target person with their eyes open. The GAN, thus not only learns what eyes should go on the person, but also how the eyes of this particular person are shaped, coloured, and so on, which gave quite realistic results.
Discover the latest Business News, Sensex, and Nifty updates. Obtain Personal Finance insights, tax queries, and expert opinions on Moneycontrol or download the Moneycontrol App to stay updated!
Find the best of Al News in one place, specially curated for you every weekend.
Stay on top of the latest tech trends and biggest startup news.