Community Center | Not monitored
Tag not monitored by Microsoft.
DALL-E image
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(128, 8%, 22%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3EA%3C/text%3E%3C/svg%3E)
Anonymous
I am trying to generate few images and campaign from DALL-E , but the images of people generally have distorted face is there way I could fix it I tried prompt engineering like specifying pretty/smart face but it does not work all the time. And for generating campaign it generates campaign with gibberish sentences please help me out in getting proper campaign with the sentences I feed as prompt.
Community Center | Not monitored
{count} votes
-
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(22.400000000000002, 11%, 12%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3ERM%3C/text%3E%3C/svg%3E)
Ramr-msft 17,836 Reputation points2023-05-04T11:56:30.83+00:00 @Karishma Nanda Thanks , Can you please add more details about the sample and prompt that you are trying. One way to fix the distortion of faces in DALL-E is to do closeups. A portrait photo is typically a face in the center of the image, facing forward. That kind of consistency helps AI learn patterns and make more accurate inferences. When not specifying “portrait” or “close up” there is much more ambiguity, they can be facing upward, to the side, etc, what the background will contain etc. The less ambiguity and the more specific the image is you are suggesting to the AI, generally the cleaner the results will be.
https://www.reddit.com/r/bigsleep/comments/vcidx2/how_to_almost_completely_fix_the_distortion_of/
-
Anonymous
2023-05-09T09:27:58.2233333+00:00 I have attached one sample and prompt is"An image of happy handsome man wearing nice cloth, real image" My requirement is full image not just closeup and even with closeup and portrait there is some distortion. What would you suggest for campaign with gibberish text ? prompt "An image of campaign saying '50% off ' "
Accept answer
0 commentsNo commentsReport a concern
Sign in to comment
2 answers
Sort by: Most helpful
-
-
Anonymous
2023-05-04T12:34:11.4666667+00:00 I have attached one sample and prompt is"An image of happy handsome man wearing nice cloth, real image" My requirement is full image not just closeup and even with closeup and portrait there is some distortion. What would you suggest for campaign with gibberish text ? prompt "An image of campaign saying '50% off ' "