New Software Accurately Predicts How Children Will Look As They Grow Up

Predicting the future, grown-up face of someone who is a child at present is a complex art. A new software seems to have made significant progress in accurately predicting the face changes of a child over the years.

Face prediction software

Predicting the changes in a person’s face is a critical part of identifying such children who are lost at an early age. Since human face changes significantly as one ages, it is often hard to identify a grown-up person from his or her childhood photographs. Due to extreme changes in face, the person in the childhood photograph and the grown-up may seem two entirely different people, even if they are the same.

However, a new software seems very helpful in this regard. This is certainly not the first time scientists have tried to automate the process, although the accuracy exhibited by the software is unmatched. It was created by three researchers, hailing from the University of Washington and Google.

Real photos (right) vs software-generated photos (left)
Real photos (right) vs software-generated photos (left)

In an attempt to be able to read human faces and their evolution over the years in detail, the trio considered the photos of 1500 people from every age group. In total, the team’s research analyzed a whopping 40,000 photos of people aged from 0 to 100. They drew on different sources to gather these photos and then analyzed them, building a prediction algorithm as they did so.

The final algorithm that the team wrote can actually calculate and visually present how a person will look in later years by taking into account factors such as the difference between the photos of people at an early age, say A, and their looks at a later age, say B. In predicting the evolution of a person’s face, all the software now needs is a single photo.

If you furnish it with the photo of a child, for example, it will churn out a whole series of photos which will show the said child’s face changes from his current picture all the way to age 80. So far, the software’s results have proved stunningly accurate, an example of which is posted above. However, it needs further improvements and tweaks before it can become a definitive tool that can be used by the authorities to identify missing children.

You can check the out the details of the work by the research team here.

Source: Illumination-Aware Age Progression
Courtesy: Pop Science

[ttjad keyword=”atnt-contract-phone”]


Salman Latif is a software engineer with a specific interest in social media, big data and real-world solutions using the two.Other than that, he is a bit of a gypsy. He also writes in his own blog. You can find him on Google+ and Twitter .

Leave a Reply