Laboratory Class 11

In this class, we will study differentiable rendering from a rasterization perspective.

Our goals are to understand the following:

• Why differentiable renderer is important.
• Which operations of the rendering pipeline are non-differentiable
• Strategies to implement a differentiable renderer

👉🏾 Download the slides here

References:

1. PyTorch3D: Getting Started with Renderer
2. OpenDR: An Approximate Differentiable Renderer paper
3. Neural 3D Mesh Renderer - paper
4. Soft Rasterizer: A Differentiable Renderer for Image-based 3D Reasoning - paper
5. Pulsar: Efficient Sphere-based Neural Rendering paper