Mass Randomized Real-Time Rendering of Scene (NVIDIA Collab)

Type: 
PR
Persons: 
1
Workgroup: 

Description

We are researching the use of efficient artificial inteligences to optimize the use of real-time rendering resources. In particular, we are looking into optimizing Variable Rate Shading, and collaborating with researchers at NVIDIA, authors of the current state of the art in the topic, to better do so.

We are looking with urgency for a student that could create a program to generate training data.

Tasks

The student will generate training data by modifying Falcor (NVIDIA's research real-time rendering framework). The student will create a program that will:

  1. Load a falcor complex scene.
  2. Compute its bounding volume and voxelize it.
  3. Perform flood-fill from a manually defined starting voxel.
  4. Randomly pick camera locations within the flooded voxels.
  5. Perform deferred physically-based shading and render results to image files.

Requirements

  • Knowledge of English language (source code comments and final report should be in English).
  • Knowledge of C++ and Vulkan or DirectX 12.
  • Available to start ASAP.

Contact

For more information please contact Joao Afonso Cardoso (jaliborc@cg.tuwien.ac.at).