The term Non-Photorealistic Rendering (NPR) refers to a family of rendering techniques that produce stylized, simplified or abstract images based on 3D geometry. Unlike photorealistic rendering, their purpose is not the simulation of the real world in as much detail as possible, but the abstraction of objects within a specific context. Examples of this are visual simplification of architecture and highlighting of geometric features such as edges and corners, the creation of blueprints or exploded view diagrams for models of complex mechanical parts, and the transformation of 3D worlds into distinct works of art with a visually consistent style.
One of the main challenges in creating NPR effects is to design the algorithms such that the stylizations are consistent between different viewing perspectives to allow for stereoscopic perception and temporal plausibility, for example when models are animated.
In this project, we will review the NPR literature, with a focus on whether existing algorithms are suitable in the context of a variety of virtual reality applications. In order to support a wide range of operating systems, such as Windows, Linux and Android, as well as different display devices such as head-mounted displays and projection-based systems, the techniques will be implemented entirely with Unity, using hand-crafted shader pipelines and Unity's abstract graphics API. In this way, the same code base can be used to create NPR rendering assets that work on any device without major changes. Depending on the size of the project group, the results of the initial research phase, and your individual interests, we will explore illustrative rendering techniques in different scenarios using a combination of architectural models, volumetric datasets, realistic avatar representations, procedural geometry, and skeletal animated models.
Do you want to learn about and create compelling NPR effects for virtual reality applications? Do you want to dive into illustrative rendering and graphics programming within Unity? Do you have at least a coarse understanding of a rasterization-based rendering pipeline? If you answered "yes" to the questions above, we would look forward to welcoming you in our project!
Kalnins, R. D., Davidson, P. L., Markosian, L., Finkelstein, A. (2003):
Coherent Stylized Silhouettes, ACM Transaction on Graphics (Proc. SIGGRAPH): 22, pp. 856-861.
Nienhaus, M. (2005):
Real-Time Non-Photorealistic Rendering Techniques For Illustrating 3D Scenes And Their Dynamics, Dissertation, Potsdam.
Hegde, S., Gatzidis, C., Tian, F. (2012):
Painterly rendering techniques: a state-of-the-art review of current approaches, Computer Animation & Virtual Worlds, 24: pp. 43-64.
Lee, Y., Kim, Y., Kang, H., Lee, S. (2013):
Binocular Depth Perception of Stereoscopic 3D Line Drawings, Proceedings of the ACM Symposium on Applied Perception, Dublin, Ireland, pp. 31-34.