Connect with us


Google and MIT are Developing Self-Correcting Photo Technology




The average person has approximately 630 photos stored on their smartphone, according to research by Gigaom. Unfortunately, many of these photos are discarded because they contain visual imperfections like blurring, red eyes, discoloration and glare. While users can always touch up their digital photos using a graphics editing program, researchers at Google and Massachusetts Institute of Technology (MIT) are developing an easier solution.

According to Android Authority, Google is working with MIT to develop a new smart-learning algorithm to touch up photos in real time. When the algorithm detects a flaw, it will automatically correct the problem; thus, eliminating the need for smartphone users to manually touch up their photos.

There are currently several ways for smartphones users to touch up their photos, including the use of graphics editing programs and apps. Some apps, for instance, will automatically touch up a user’s photos. The problem with this method, however, is that it typically either applies the change to all photos when activated, or it requires the user to manually activate the touch-up feature. The algorithm being developed by Google and MIT will only touch up photos that contain visual flaws — and will perform this process while the user is preparing to take the picture.

So, how does Google and MIT plan to develop this self-correcting photo algorithm? The team used a sample of 5,000 images to train their algorithm what imperfections to look for. According to Google’s Jon Barron, the algorithm can be used on smartphones to improve photo-taking features without sacrificing battery power. To make the changes, the algorithm uses a low-resolution version of the photo, which subsequently saves up to 90% processing power.

There’s still no word on when Google will release its new self-correcting photo technology, but industry experts believe this feature will soon find its way to the Android operating system.

Continue Reading
Click to comment

Leave a Reply

Your email address will not be published.

17 − seven =