diff options
Diffstat (limited to 'doc/CNN_V2.md')
| -rw-r--r-- | doc/CNN_V2.md | 10 |
1 files changed, 5 insertions, 5 deletions
diff --git a/doc/CNN_V2.md b/doc/CNN_V2.md index 2913498..78854ce 100644 --- a/doc/CNN_V2.md +++ b/doc/CNN_V2.md @@ -122,12 +122,12 @@ let uv_x = coord.x / resolution.x; // Horizontal position [0,1] let uv_y = coord.y / resolution.y; // Vertical position [0,1] // Slot 6: Multi-frequency position encoding -let sin10_x = sin(10.0 * uv_x); // Periodic feature (frequency=10) +let sin20_y = sin(20.0 * uv_y); // Periodic feature (frequency=20, vertical) // Slot 7: Bias dimension (always 1.0) let bias = 1.0; // Learned bias per output channel -// Packed storage: [p0, p1, p2, p3, uv.x, uv.y, sin(10*uv.x), 1.0] +// Packed storage: [p0, p1, p2, p3, uv.x, uv.y, sin(20*uv.y), 1.0] ``` ### Feature Rationale @@ -136,7 +136,7 @@ let bias = 1.0; // Learned bias per output channel |---------|-----------|---------|----------| | p0-p3 | 4D | Parametric auxiliary features (mips, gradients, etc.) | Essential | | UV coords | 2D | Spatial position awareness | Essential | -| sin(10\*uv.x) | 1D | Periodic position encoding | Medium | +| sin(20\*uv.y) | 1D | Periodic position encoding (vertical) | Medium | | Bias | 1D | Learned bias (standard NN) | Essential | **Note:** Input image RGBD (mip 0) fed only to Layer 0. Subsequent layers see static features + previous layer output. @@ -149,8 +149,8 @@ let bias = 1.0; // Learned bias per output channel ### Future Feature Extensions -**Option: Replace sin(10\*uv.x) with:** -- `sin(20*uv.x)` - Higher frequency encoding +**Option: Additional encodings:** +- `sin(40*uv.y)` - Higher frequency encoding - `gray_mip1` - Multi-scale luminance - `dx`, `dy` - Sobel gradients - `variance` - Local texture measure |