Topaz Labs has released a new version of its Gigapixel AI software, an application that allows photographers and graphic artists to enlarge and improve the resolution of their images far beyond what’s possible with traditional techniques such as bilinear or bicubic resampling. The new v5.8 update enables a number of improvements and fixes that include improved face refinement for more realistic looking faces, reduced memory usage, and, most importantly, better image quality with Windows GPUs. Gigapixel AI works much faster when its AI Processor is set to GPU, but in previous versions, its output quality was generally inferior compared to CPU.
Gigapixel AI v5.8 – Improved face refinement, reduced memory usage, better image quality with Windows GPU, and lots of stability fixes (Topaz Labs)
- Improved face refinement pipeline – Gigapixel AI will now do a much better job detecting human faces more often and delivering more realistic results. Tiny background faces especially should see a noticeable improvement in quality.
- Significant image quality improvements with Windows GPU – You’ll notice much better image quality across all models in Gigapixel AI when using GPU-powered Windows machines.
- Reduced memory usage – We’ve streamlined how we cache and utilize image data through the entire editing pipeline, resulting in less memory usage and faster importing and exporting of image files.
- Lots of usability improvements and bug fixes – We’ve improved the post-processing resize algorithm, optimized how we compress JPG, PNG, and TIFF files when exporting, and squashed a bunch of annoying bugs.