ComfyUI Node
CV Quality BRISQUE Features
The 36 natural-scene-statistics features BRISQUE scores with (cv2.quality.QualityBRISQUE_computeFeatures) - mean/variance/shape of locally normalized luminance at two scales, plus its pairwise products in four directions. Needs NO model file. As a generic descriptor of how 'natural' an image's statistics are, feed it to 'CV Train Classifier' to learn your OWN quality or defect classifier instead of accepting the LIVE-trained SVM's notion of quality.
CV Quality BRISQUE Features
- images
- features
- count
Categoryimage/CV/quality
Inputs (1)
| Name | Type | Default | Description |
|---|---|---|---|
| images | NPARRAY,IMAGE,MASK | Image(s) to describe. One feature row per frame. Accepts a ComfyUI IMAGE/MASK directly (frame 0 of a batch) or an NPARRAY. Arithmetic ops (add, multiply, etc.) process the full IMAGE batch when both inputs have the same batch size. |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| features | NPARRAY | float32 array of shape (N, 36) - one row per frame, ready as a sample matrix for 'CV Train Classifier' / 'CV Stack Feature Classes'. |
| count | INT | Number of feature rows (= batch size). |