Image Regeneration with Human and AI collaboration

We are working to make a product that regenerates more detailed and extensive images with the same context given as an input image by the user. Image is inserted as a reference image to the model that generates editable captions (five each). The caption selected by the human is used as a final decision for the machine learning model (Stable Diffusion), which uses that textual data as an input to regenerate set of images. It is a web based application inheriting all the features of human in the loop approach and advance AI model. User is given the opportunity to download the desired images to its personal computer or re-diffused for the further regeneration. Its an hybrid model that rely on both AI and Human interaction.