Google has released a new open source image compression algorithm called Guetzli. The latest development is claimed to reduce the size of JPEG images by 35 per cent.
Guetzli creates smaller file sizes without impacting the quality. The algorithm patches the differences between psychovisual modeling of the JPEG format by employing a search algorithm. This process approximates color perception and visual masking in a detailed way to overcome the differences.
“Guetzli specifically targets the quantisation stage in which the more visual quality loss is introduced, the smaller the resulting file,” Google’s research team, comprises Robert Obryk and Jyrki Alakuijala, writes in a blog post.
Usually, image compression algorithms use simple color transforms, and cosine transforms to overcome pixel differences. But the huge advantage that Guetzli has is the output files that are compatible with all existing browsers, JPEG readers and image processing programs.
Google has published the pre- and post-compression screenshots of the images to visually represent the differences.
The screenshots show that the compressed images have less ringing artefacts than the libjeg version that reduces the image size to some extent.
A cookie to save your storage
Guetzli means cookie in Swiss German — subjectively aims to create smaller size of images without impacting the pixel quality.
The only disadvantage of Guetzli is the time taken to compress the images. It takes slightly more time than most of the existing image compression algorithms.
You can access Guetzli from a GitHub repository. The online repository includes all the necessary documentation to help you deploy the algorithm for your next project.