Learning for 3D Homeworks
Homeworks in CMU 16-825: Learning for 3D Vision.
Assignment1: Rendering Basics with PyTorch3D
Practicing with Cameras
Rendering Your First Mesh
Practicing with Cameras
360-degree Renders
Re-creating the Dolly Zoom
Practicing with Meshes
Constructing a Tetrahedron
Constructing a Cube
Re-texturing a Mesh
Rendering Generic 3D Representations
Rendering Point Clouds from RGB-D Images
Parametric Functions
Implicit Surfaces
Sampling Points on Meshes
Assignment2: Single View to 3D
Exploring Loss Functions
Fitting A Voxel Grid
Fitting A Point Cloud
Fitting A Mesh
Reconstructing 3D from Single View
Image To Voxel Grid
Image To Point Cloud
Image To Mesh
Assignment3: Volume Rendering, Neural Radiance Fields, Neural Surfaces
Neural Volume Rendering
Volume Rendering
Optimizing A Basic Implicit Volume
Optimizing A NeRF
Neural Surface Rendering
Sphere Tracing
Optimizing A Neural SDF
VolSDF
Assignment4: 3D Gaussian Splatting and Diffusion Guided Optimization
3D Gaussian Splatting
3D Gaussian Rasterization
Training 3D Gaussian Representations
Rendering Using Spherical Harmonics
Training A Harder Scene
Diffusion-guided Optimization
SDS Loss + Image Optimization
Texture Map Optimization for Mesh
NeRF Optimization
I met some problem compiling gridencoder and failed to train the model. Maybe finish this part in the future…
Assignment5: Point Cloud Processing
PointNet for Segmentation
PointNet++ for Segmentation
Seems that PointNet++ does not improve the performance. Maybe need more effort fine tuning…
