More efficiency for AI art

Can https://perchance.org/ have an ability for use an image (like a file of traditional art which can be uploaded) to generate image directly (without turn into prompt text and then making image), this would be more effective to its machine learning and also for users. We want to help as volunteers!