1.5 KiB
1.5 KiB
Multi Mask Math Ops - Binary Mask Operations
Node Purpose: Mask Math Ops performs per-pixel binary-like operations between two masks: or, and, subtract (a-b), subtract (b-a), and xor. Handles batch cycling and resizes masks to match when sizes differ.
Inputs
| Name | Port | Type | Default | Range | Description |
|---|---|---|---|---|---|
mask_1 |
- | MASK | - | - | First mask batch. |
mask_2 |
- | MASK | - | - | Second mask batch; resized to mask_1 size when needed. |
operation |
- | COMBO | or |
or / and / subtract (a-b) / subtract (b-a) / xor |
Operation type. |
Outputs
| Name | Type | Description |
|---|---|---|
mask |
MASK | Resulting mask batch. |
Features
- Batch cycling: aligns differing batch sizes by indexing modulo per item.
- Size alignment: resizes
mask_2tomask_1with Lanczos when shapes differ. - Operations:
and: min of arrays.or: max of arrays.subtract (a-b):clip(a-b, 0, 1).subtract (b-a):clip(b-a, 0, 1).xor: absolute difference.
Typical Usage
- Combine masks: union/intersection of regions for composite selection.
- Cutouts: subtract one region from another to refine selections.
- Edge emphasis:
xorhighlights differences between masks.
Notes & Tips
- Inputs are treated as float
[0,1]arrays internally; operations are per-channel on single-channel masks. - Ensure consistent semantic meaning of mask intensity (white=selected) across inputs before combining.