Image Super-Resolution
The cluster focuses on discussions about super-resolution and AI-based image upscaling techniques, comparing their performance to traditional methods like bicubic interpolation, and addressing issues like information loss or hallucinations in results.
Activity Over Time
Top Contributors
Keywords
Sample Comments
Was hoping they used super-resolution magic to scale up the image.
Photoshop also has super resolution
The upscaled version looks worse to me. Seems like it's lost information, not gained.
Have you thought of running super resolution on the images?
Almost like super resolution techniques?
What's wrong with upscaling?
You can't really use machine learning based image upscaling for that type of work
Isn't this an issue with the algorithm?Can someone try and see how it would perform if you simply upscale the image using normal bicubic interpolation? And if it performs much better, I feel like that should be a preprocessing option to scale up the image since it seems to do so poorly on small resolutions.
You're mistaken and the original experiment does not distinguish between classic edge aware upscaling/super resolution vs more problematic replacement
Its not upscaling. Its taking what it knows about other non ST pictures and creating new texture and information.