The debate surrounding the extent of image manipulation continues. Yet, artificial intelligence (AI) has become an integral part of computational photography and even extends into the post-photography process. Google, a significant player, introduced Magic Eraser with the Pixel 6 two years ago, which efficiently removed unwanted objects from photos.
AI’s leap into photo editing
However, object removal is just a tip of the iceberg in photo editing. Google’s forthcoming Magic Editor promises a lot more. The company has recently announced several updates to Pixel’s inbuilt photo editor app, aiming to combine generative AI and semantic segmentation. Yet, skepticism remains due to Google’s past record of making announcements but not delivering. Plus, why wait for Google when Adobe’s Photoshop on the desktop is already providing these capabilities?
Testing Photoshop’s AI
Adobe recently released Photoshop Beta with generative AI features, currently available only on the desktop version. Does it live up to the hype? After testing it across various scenarios, here are the results.
Considering AI for editing | Votes |
---|---|
Already tried Photoshop beta | %54 |
Prefer not to alter images | %46 |
Generative AI in Photoshop enables expansion beyond the original photo. In addition, it can be used to create humorous modifications or memes.
Photoshop AI vs. Google’s Magic Eraser
Where Google’s Magic Eraser falls short in object removal and background generation due to on-device ML limitations, Photoshop’s generative AI provides cleaner results.
Limitations of AI editing
Despite the promise, AI-based editing still has limitations. For instance, it struggles with matching exposure and contrast levels in a photograph, implying that manual intervention is still needed for achieving photorealism.
The power and problem of altering reality
Photoshop’s generative AI can completely alter a photograph by replacing the background and adding new elements. However, the result is not flawless, indicating the need for further refinement.
What are your thoughts on this topic? Please share your views in the comments section below!
{{user}} {{datetime}}
{{text}}