Beta: You might’ve used many age detecting tools, what about a tool that makes you look older or younger? A team of researchers from Orange Labs, France, has used two machine learning techniques to design a system that does this process with 80% accuracy. Called Age Conditional Generative Adversarial Network, this computer system can be used to identify the people who’ve been missing for a long time.
Do you remember how badly Microsoft’s age detecting robot failed? Having said that, ever wondered how you’ll look in the next 30 or 40 years? There are some techniques available to do this for you. While many of them are unable to retain the identity of the faces in the process, others are very expensive and time-consuming. Now, some researchers have used machine learning to do this job easily.
2 deep learning machines — Face generator and face discriminator
The approach used by Antipov’s team uses two machine learning systems that involve a face generator and a face discriminator.
These machines were trailed using 5,000 faces in the age groups 0-18, 19- 29, 30-39, 40-49, 50-59, and 60+ years old. These faces were picked from IMDb and Wikipedia. The machine was able to learn the signature features of each group, which were used to apply on other faces.
After this, face discriminator system studies the synthetically generated faces and tells whether the original identity can still be picked out.
In the tests carried out by the researchers, their system spotted the correctly aged faces about 80% of the time, which is higher than 50% of the time for other processes.
But, what’s the use of such system? The researchers say that their technique can be employed to be used to find the people who have been missing for many years. Well, let’s see how effectively this technique is used in real life situations in near future.
Did you find this face aging technique interesting? What could be its future applications? Don’t forget to share your views.