I think what you are looking for is image-to-image AI, and Perchance doesn’t have that. I would suggest looking into creating your own local Stable Diffusion setup.

If you want to know the additional text applied to your prompt when using the t2i framework, see this styles page.