The AI algorythm is struggling to understand the prompt to create an image/ art/ graphic in the Indian context. This becomes even messier when it comes to creating imagery about religion, god, faith and belief etc. It would be a good idea to allow the user to add a reference image along with a written prompt for the AI algorythm to understand the prompt better. Currently, one can use the reference image from the four results generted from text to image format. However, there's no mechanism to add a reference image from the user along with the text prompt.
Please help take this forward.
I was trying to crate an artwork of Pandurang, a Hindu deity that is an incarnation of Vishnu, the creator of the universe. I tried different prompts. Firefly got one very good image. However, when I used it as a reference point, the results could not be replicated at all.
Some images generated are not accurate/full in nature. Example: prompt used for image generation of Indian/Hindu Deties contain broken hands, fingers etc.
6
Replies
AdChoices