PixelBatch<T>
Models & Types · AiDotNet.NeuralRadianceFields.Data
A batch of pixels sampled from one or more ImageView views — the target
side of image-space radiance-field training (#1834). Each sampled pixel carries a ray
origin + direction (so the model can render exactly that pixel through volume rendering)
and the ground-truth RGB the loss compares against.
Constructors
| Constructor | Summary |
|---|---|
PixelBatch(Tensor<T> rayOrigins, Tensor<T> rayDirections, Tensor<T> targetColors) |
Properties
| Property | Summary |
|---|---|
Tensor<T> RayDirections { get; } |
Ray directions (unit vectors) in world coordinates [N, 3]. |
Tensor<T> RayOrigins { get; } |
Ray origins in world coordinates [N, 3]. |
Tensor<T> TargetColors { get; } |
Ground-truth pixel colors [N, 3] — the target the photometric loss compares the model’s rendered output against. |
int Count { get; } |
Number of pixels / rays in this batch. |