ControlNetConditionType
Enums · AiDotNet.Diffusion.Control
Specifies the type of conditioning input for multi-control ControlNet composition.
For Beginners
When using multiple ControlNet conditions at once (e.g., both a sketch and a depth map), the model needs to know what each input represents so it can combine them properly. This enum labels each input with its type, like putting labels on different ingredients before mixing them together.
How It Works
ControlNet models can accept multiple simultaneous control conditions (e.g., Canny edges + depth map). This enum identifies the semantic type of each conditioning input to enable proper routing, weighting, and compositing of multiple control signals.
Fields
| Field | Summary |
|---|---|
ControlNetConditionType Brightness |
Brightness/luminance map for lighting control. |
ControlNetConditionType CannyEdge |
Canny edge detection map for structural guidance. |
ControlNetConditionType ColorMap |
Color map for per-pixel color guidance. |
ControlNetConditionType ColorPalette |
Color palette for palette-guided generation. |
ControlNetConditionType ContentShuffle |
Content shuffle for structure-preserving randomization. |
ControlNetConditionType DWPose |
DWPose whole-body keypoint detection. |
ControlNetConditionType DepthMap |
Monocular depth estimation for spatial layout. |
ControlNetConditionType FaceID |
FaceID embedding for identity-preserving generation. |
ControlNetConditionType HED |
HED (Holistically-Nested Edge Detection) soft edges. |
ControlNetConditionType InpaintMask |
Binary inpainting mask (white = inpaint region). |
ControlNetConditionType InstanceSegmentation |
Instance segmentation with individual object masks. |
ControlNetConditionType LineArt |
Line art or clean sketch drawing. |
ControlNetConditionType MLSD |
MLSD (Mobile Line Segment Detection) straight lines. |
ControlNetConditionType MediaPipeFace |
MediaPipe face mesh landmarks. |
ControlNetConditionType NormalMap |
Surface normal map for lighting and geometry. |
ControlNetConditionType OpenPose |
OpenPose body and hand keypoints. |
ControlNetConditionType QRCode |
QR code pattern for embedding readable codes. |
ControlNetConditionType Recolor |
Recoloring target for guided color transfer. |
ControlNetConditionType ReferenceImage |
Reference image features (no explicit spatial map). |
ControlNetConditionType SAMSegmentation |
SAM (Segment Anything) segmentation mask. |
ControlNetConditionType Scribble |
User-drawn scribble or sketch. |
ControlNetConditionType SemanticSegmentation |
Semantic segmentation label map. |
ControlNetConditionType SoftEdge |
SoftEdge detection (PiDiNet or similar). |
ControlNetConditionType StyleReference |
Style reference for style-aligned generation. |
ControlNetConditionType Tile |
Tile image for detail-preserving upscaling. |
int value__ |