Pixelmator today released an update to its Pixelmator Pro Avalon image editing software. The new ML Super Resolution feature increases the resolution of images while maintaining sharpness and levels of detail, making (to a degree) the ‘enhance image’ operation in detective shows a reality.
Compared to a simple nearest neighbor upscaling algorithm the Pixelmator is capable of achieving much better results thanks to machine learning technology. It takes into account the actual content of every image, attempts to recognize edges, patterns, and textures, and recreates detail based on Pixelmator’s dataset, in a similar way to Google’s RAISR technology.
Despite the complex technology Pixelmator says the feature is very fast and takes a few seconds to upscale an image on the latest Mac computers thanks the use of Core ML 3 and multi-GPU support. It’s also entirely on-device, with the trained model integrated in the Pixelmator Pro app package. This means all processing is done locally and no data is sent to any online servers.
Additionally, the tool is capable of removing noise and JPG artifacts including compression artifacts. To do its magic the software requires macOS High Sierra and a Metal-compatible graphics card.
Pixelmator Pro 1.5 Avalon is available from the Mac App Store as a free update for existing users or for $39.99 for new customers. There is also an upgrade discount for owners of the original Pixelmator software.
More information and some interactive comparisons with other upscaling methods are available on the Pixelmator website.
Author: Go to Source