Researchers Resort to Neural Network for DSLR-Quality Images sans Dual Cam

November 1, 2017, By Sanjeev Ramachandran

Smartphone camera have been one of the prime factors that depend the success of a smartphones. Buyers are being offered broader options these days with the influx of dual-cams and quad-cams, not to forget the very own non-dual software-tweaked camera prowess of Google Pixel. No matter what the technology is, every vendor tries to boast its camera to be of DSLR-alike, which is what attracts the camera savvy buyers.

But if that truly is the sole factor, you might not want to depend on the smartphone camera anymore. A team of young researchers have now come up with a feature to help users get DSLR-alike quality on their already captured images. The technology relies on neural network to deliver DSLR-quality images which is processed by realizing how the image should look like.

The output is obtained after a serious analysis of the image in the neural network. It’s not quite DSLR-like like its being touted. Yet, this is some technology to look forward to for all the device manufacturers. You can check it here.

One interesting feature is that not all photos go through the same process to get DSLR-like output. Some may have its brightness upped, while some may have additional tints layered for the final output. That’s the result of its analysis by an AI brain that matches the photos so as to render it with the best available filters. Besides, it could also add details that are missing from the original photo, including ability to recover brightness from low-light photos.

The work is still developing, and the researchers are still on the process of adding more features. So we might not see the technology readily making it to the smartphones. But it would be interesting if smartphone manufacturers can absorb the technology and better their existing cameras.

© 2008-2012 DeviceMag.com - All rights reserved | Privacy Policy