he Pixelmator Team recently released an update to Pixelmator Pro, adding a new Core ML-powered feature that intelligently matches the colors of different photos or objects using a machine learning algorithm trained on millions of professional photos. The update also brings a new Free Transform feature to freely resize, distort, and transform layers, and includes a number of additional improvements and fixes.
“With Pixelmator Pro, our goal is to create the most powerful image editor that absolutely anyone can use,” said Saulius Dailide, one of the founders of the Pixelmator Team. “Our extensive collection of color adjustments gives users all the tools they need to bring out the best in any photo. And ML Match Colors makes these tools accessible even to complete beginners who want to quickly match the look of one photo to another.”
The new ML Match Colors feature lets anyone match the look of one photo to another in an instant. And it couldn’t be simpler to use. To match the style of a photo, users simply drag and drop images onto the Color Adjustments pane. ML Match Colors then analyzes the contents of both the source and target photo to identify any objects, detects their colors and lighting, and applies a range of color adjustments to the target photo. 37 individual adjustments in total, such as Brightness, Shadows, Highlights, White Balance, and Selective Color, are used for color matching.
ML Match Colors is also available in the Edit menu as Paste and ML Match Colors, letting users automatically color match any objects when pasting them into a new image. ML Match Colors has been trained on 20 million professional photographs and is powered by Core ML, delivering unmatched image analysis speed and performance.
Another headline feature in the update is the new Free Transform tool, which adds a number of different ways to transform objects, including the ability to skew, distort, and change the perspective of any layer. Along with perspective transformations, the Free Transform tool lets users resize individual layers using 3 specialized layer scaling algorithms – Bilinear, useful for photographic images, Lanczos, for images with small and intricate details, and Nearest Neighbor, which is most often used for pixel art.