Easy Artistry is a tool intended to aid small artists and indie studios in developing larger projects using AI image generation. The tool will consist of a desktop application that uses artist-centric features to allow image generation engines to be able to iterate on, modify, and aid in creation of artists orginal works. Easy Artistry is not intended to replace human artists, but instead to expand the scope of what an already skilled artist can acomplish with a smaller amount of time and budget. Our hope is to allow individuals to create animations, comics, and concept art, and small studios to create films.
Our goals this semester is to build a prototype desktop app, with our main features in place. Primarily, we want to be able to generate and modify images with StableDiffusion, in a way that prevents direct publication. Our stretch goals (which are looking more achievable given the unexpectedly large size of our team!) are to begin allowing users to pay the bare minimum required to generate the images over the internet, allow users to change the image generation model to one of their choice, and to begin training our own, more ethical, image generation model.
Stack: Our current plan is to use python for both front and back end, and StableDiffusion as our starter model. However, this is subject to change as our needs change, and we learn more.
Roadmap:
- Get a basic framework program running. Feb. 15
- Generate images locally using StableDiffusion. Feb 15
- Generate images in the program. Feb 30
- Implement Invisible watermark
- Implement presets.
- Implement API support.
- Begin Prototyping iteration.
- Refine iteration tools.
- Refine UI and begin branding.
- Polish website, add download link.
- Begin testing online image generation (possibly using AWS).
- Begin testing other engines.
- Begin Researching our own model.
- Start social media campaign to collect data ethically.