Nodes/Nodes-/Face Composite
ComfyUI Node

Face Composite

Composites refined face images back onto original video using bounding box coordinates. Handles frames without faces (bbox at 0,0 with original size) by passing through unchanged.

By systms-ai·Created 6 months ago·Updated 6 months ago· 0
Face Composite
  • original_images
  • face_images
  • face_bboxes
  • composited_images
blend_modefeather
feather_amount5
CategoryWanAnimatePreprocess

Inputs (5)

NameTypeDefaultDescription
original_imagesIMAGEOriginal video frames before face extraction
face_imagesIMAGERefined face images to composite back
face_bboxesBBOXFace bounding boxes from Pose and Face Detection node
blend_modeoptCOMBOfeatherHow to blend refined faces back onto original
feather_amountoptINT50–200Pixels to feather at edge (only for feather mode)

Outputs (1)

NameTypeDescription
composited_imagesIMAGE