Ground Truth Renderer

Features

Loads and renders different subsets of point clouds
View the point cloud in different modes
- Splats: high quality splat rendering of the whole point cloud
- Points: high quality point rendering of the whole point cloud
- Neural Network: renders neural network output and loss channel comparison
Compare to a sparse subset of the point cloud with a different sampling rate
HDF5 dataset generation containing color, normal and depth images of the point cloud
Blending the colors of overlapping splats
Phong Lighting

Neural network should output the splat rendering from a sparse point rendering input
The neural network must be loaded from a selected .pt file with Load Model
A description of the input/output channels of the network must be loaded from a .txt file with Load Description
Each entry consists of:
- String: Name of the channel (render mode)
- Int: Dimension of channel
- String: Identifying if the channel is input (inp) or output (tar)
- String: Transformation keywords e.g. normalization
- Int: Offset of this channel from the start channel
Example for a simple description file:

[['PointsSparseColor', 3, 'inp', 'normalize', 0], ['SplatsColor', 3, 'tar', 'normalize', 0]]

When a Loss Function is selected
- Loss between two channels (Loss Self and Loss Target) is computed
- Screen area between these channel renderings can be controlled Loss Area

Rendering resolution must currently be a power of 2
There are two modes for the dataset generation with parameters in the HDF5 Dataset tab
- Waypoint dataset:
  - Interpolates the camera between the user set waypoints
  - Use the Advanced tab to add/remove a waypoint for the current camera perspective
  - Preview Waypoints shows a preview of the interpolation
  - Step Size controls the interpolation value between two waypoints
- Sphere dataset:
  - Sweeps the camera along a sphere around the point cloud
  - Step Size influences the amount of viewing angles (0.2 results in ~1000 angles)
  - Theta and Phi Min/Max values define the subset of the sphere to be sweeped
  - Move the camera to the desired distance from the center of the point cloud before generation
Start the generation process with Generate Waypoint HDF5 Dataset or Generate Sphere HDF5 Dataset
- Make sure that the density and rendering parameters for the Sparse Splats view mode are set according to Configure the rendering parameters
- The generated file will be stored in the HDF5 directory
- After generation all the view points can be selected with the Camera Recording slider