3D Scene Reconstruction and Synthesis

I propose research into the creation of a tool that generates 3D environments given some input styling, either in the form of drawings, still images, video, text, or more.

I propose that we iterate on existing solutions in the space of object and scene generation, by pairing them together we should be able to solve our problem.

This research is useful culturally as it helps to democratise creative endeavours, an author would no longer need to outsource a video game adaptation to a third party studio, they could simply pass their book into this model. Indie movie studios could far more easily create CGI environments.