Advances in Neural Rendering

Synthesizing photo-realistic images and videos is at the heart of computer graphics has been focus decades research. Traditionally, synthetic a scene are generated using rendering algorithms such as rasterization or ray tracing, which take specifically defined representations geometry material properties input. Collectively, these inputs define actual what rendered, referred to representation (where consists one more objects). Example triangle meshes with accompanied textures (e.g., created by an artist), point clouds from depth sensor), volumetric grids CT scan), implicit surface functions truncated signed distance fields). The reconstruction observations differentiable losses known inverse rendering. Neural closely related, combines ideas classical machine learning create for synthesizing real-world observations. leap forward towards goal image video content. In recent years, we have seen immense progress in this field through hundreds publications that show different ways inject learnable components into pipeline. This state-of-the-art report on advances neural focuses methods combine principles learned 3D representations, often now representations. A key advantage they 3D-consistent design, enabling applications novel viewpoint synthesis captured scene. addition handle static scenes, cover modeling non-rigidly deforming objects...