I'm new to Adobe Stock, and many of my images are getting rejected for "quality issues." I've already upscaled them to over 4MP, but they still keep getting rejected. Is the problem the resolution, or something else?

Quality & Technical factors are much more than pixels alone. See list below.

https://helpx.adobe.com/stock/contributor/help/quality-and-technical-issues.html

• AI images should make sense visually. Check perspectives & small details. 
• Is it free of unwanted artifacts?
• Check your Histogram panel in Photoshop or Lightroom. Is the lighting balanced and coming from logical light sources? 
• Is most of the image in focus?
• Is white balance neutral?

The first image appears too dark & too blue to me. A Curves Adjustment layer in Photoshop could help correct this.

Hope that helps.

​@leocetke Here you go. I know a video would be better but there are just too many variables and potential for errors when working with AI. 
 

The following steps assume that basic editing has already been completed.

1. 1. Upscale your image. When editing with AI, it’s helpful to provide the bot with as many pixels as reasonable.
   2. Select a portion of the image that requires editing. This can be a hand, an entire face, eyes, ears, a misshapen object, hair, clothing, a computer…whatever is in need of being fixed. To make the steps that follow easier, the selection should be square and, when possible, under 6.5 inches or so in width and length. I generally use the paid version of Gemini 3, which upscales to 4K, so I’m able to reduce the size of the result rather than upscale it. (Important: try to leave the original selection in place.)
   3. As an example, let’s assume we’re using a woman’s misshapen hand or one with a missing or sixth finger. Select it, copy it, and paste it into Gemini 3’s prompt window. (If using Gemini 3 or other models within Firefly, you will need to save the image first in order to upload it.)
   4. I’d use the following prompt: “Improve the anatomy and quality of the woman’s hand, with particular attention to the fingernails. Leave everything else as is.” (Side note: It may or may not be important to tell the AI that the hand is that of a woman, but if you end up with the hand of a man, then yes, it’s important. I always call attention to the fingernails , which seems to help assure higher quality results. The “leave everything else as is” in the prompt is extremely important to help assure the AI does not make unwanted changes to the surrounding subject.
   5. If successful, save the result. If not, tweak the prompt until the AI understands exactly what you want fixed.
   6. Save the result and open it in Photoshop.
   7. Return to your original image where you first made your selection and select “New.” The size of your original selection as copied should still be in memory. Note the size in pixels or inches.
   8. Return to your AI image and select Image > Image Size… and resize it to the size of the copied selection. Copy and paste it into the selected area of the original image. If you’re lucky, everything should match up, but you aren’t quite finished.
   9. Hide the layer of the original image, then use a mask and paintbrush, or the erasure tool, to remove the edges of the top image. This helps assure there will be no noticeable hard lines around the finished image.
   10. Flatten the layer and move on to the next area or object that needs editing.

Good luck. Let me know if you have any questions.

Hey ​@leocetke 

I am going to chim in for a second. ​@daniellei4510  
Great looking images.

Bottom image
Consider dropping the exposure of the window. It’s pulling my eye away from the subject.
Delete unnecessary distractions: white bit on the tablet, the books behind the chair, the phone beside the computer.
Extend the desk so it fills the screen at the front and on the side.
Remove that gray window frame. AI is powerful. 
Remove the line on the red/ orange wall outside. 


Cheers

Nate

Any quality issue aside, this needs to be removed.