Despite significant progress in Computer Graphics, HCI and Computer Vision, creating, manipulating and displaying high-quality visual content remains reserved to a small, well-trained professional elite. We believe that this difficult access is largely due to the core theoretical foundations of Computer Graphics, which were developed for accurate and complete synthetic content.
The goal of GraphDeco is to redefine these foundations so they can simultaneously account for both traditional content as well as approximate and incomplete representations captured or produced by casual users. In particular, we plan to explore two main research directions:
- First, the design and artistic principles that guide the creation of visual content need to be automated and redesigned, facilitating professional practices and making such endeavors accessible to all.
- Second, a unified representation is required to develop new computer graphics models and algorithms capable of handling traditional synthetic content as well as inaccurate and incomplete captured data.
The GraphDeco team is a follow-up of the Reves team.