ComfyUI Node
CV Coordinate Grid
A dense meshgrid of pixel coordinates: every (x, y) on a step-spaced grid over a width x height canvas, row-major (y outer, x inner). Feed it into 'CV Train Classifier' query_features to classify EVERY grid location, then 'CV Reshape Array' the predictions back to (rows, cols) and colormap them - that renders the decision REGIONS of the classifier. Also a generic sample-position generator for remap/lookup experiments.
CV Coordinate Grid
- points
- cols
- rows
- count
◄width224►
◄height224►
◄step2►
Categoryimage/CV/ml
Inputs (3)
| Name | Type | Default | Description |
|---|---|---|---|
| width | INT | 2241–8192 | Canvas width in pixels the grid covers (x goes 0, step, 2*step, ... below width). |
| height | INT | 2241–8192 | Canvas height in pixels the grid covers (y goes 0, step, 2*step, ... below height). |
| step | INT | 21–1024 | Grid spacing in pixels. 1 = every pixel (slow to classify); 2-4 is plenty for a preview. |
Outputs (4)
| Name | Type | Description |
|---|---|---|
| points | NPARRAY | Nx1x2 float32 (x, y) grid coordinates, row-major: N = rows * cols. Feeds query_features of 'OpenCV Train Classifier'. |
| cols | INT | Grid columns = number of x positions. Use as the cols of 'CV Reshape Array' on the predictions. |
| rows | INT | Grid rows = number of y positions. Use as the rows of 'CV Reshape Array' on the predictions. |
| count | INT | rows * cols. |