Global EditionASIA 中文双语Français
Opinion
Home / Opinion / Opinion Line

AI comes to the rescue of vague images

By Zhang Zhouxiang | China Daily | Updated: 2019-04-30 07:53
Share
Share - WeChat
[Photo/IC]

SOME PHOTOGRAPHS of famous actress Bridgitte Lin Ching Hsia taken in the 1980s have been "repaired", that is, transformed from blurred into high-density images, and spread online. China Daily's reporter Zhang Zhouxiang explains the artificial intelligence (AI) technology behind the miracle:

We can make a clear photo vague by lowering its density using any image processing software. But until now there was no software which could do the opposite: make a vague image clear.

Most image-processing techniques involve treating the image as a two-dimensional signal and applying signal-processing techniques to it. The higher the density of an image the more information it contains. Which means the software reduces information to make a clear photo vague. And since a vague photo, taken by a camera with fewer pixels, contains less information, the software needs additional information to make it clear. Yet it cannot create the information out of nothing.

Information is lost from the object when a low-density camera takes its photos, and that information cannot possibly be recovered.

Thanks to its fast development, AI technology can "repair" a vague photo in several ways, but the basic science behind it is the same: programmers let the AI scan the old photos to understand the hidden rules of the colors and changing light beams, and then put the pixels on the image to increase its density.

AI does not "recover" the lost information but "guesses" the rules of the color and light, and imitates the lost pixels of the photo. The better algorithms an image processing software has the more it learns from the image, and the more familiar it is with the rules the better it can imitate the "real" photo. The transformation of Lin's photos shows the technology has become quite mature.

The technology can be applied in other sectors, too. For example, in emergencies such as a forest fire, firefighters often don't know what is the condition inside the forest because the photos taken by drones are not so clear. By using AI technology, the firefighters can gather more information and act accordingly.

Most Viewed in 24 Hours
Top
BACK TO THE TOP
English
Copyright 1995 - . All rights reserved. The content (including but not limited to text, photo, multimedia information, etc) published in this site belongs to China Daily Information Co (CDIC). Without written authorization from CDIC, such content shall not be republished or used in any form. Note: Browsers with 1024*768 or higher resolution are suggested for this site.
License for publishing multimedia online 0108263

Registration Number: 130349
FOLLOW US