Skip to main content
The SDPoseFaceBBoxes node processes pose keypoint data to detect and generate bounding boxes around human faces. It analyzes the 2D face keypoints for each person in a frame, calculates a bounding box based on those points, and can adjust the box’s size and shape. The resulting bounding boxes are formatted to be compatible with other nodes in the SDPose workflow, such as the SDPoseKeypointExtractor.

​
Inputs

ParameterDescriptionData TypeRequiredRange
keypointsThe pose keypoint data containing information about detected people and their body/face landmarks per frame.POSE_KEYPOINTYes-
scaleMultiplier for the bounding box area around each detected face. A larger value creates a larger box. (default: 1.5)FLOATNo1.0 - 10.0
force_squareExpand the shorter bbox axis so the crop region is always square. (default: True)BOOLEANNo-
Note: The keypoints input must be in the specific format produced by nodes like SDPoseKeypointExtractor, containing canvas_height, canvas_width, and people data with face_keypoints_2d for each person.

​
Outputs

Output NameDescriptionData Type
bboxesA list of face bounding boxes for each frame. Each bounding box is defined by its top-left coordinates (x, y), width, and height. This output is compatible with the bboxes input of the SDPoseKeypointExtractor node.BOUNDINGBOX
This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHub
Source fingerprint (SHA-256): dc68640f46afcba4c10a907fd9467671d508cd5bac7670b35a038d159f9d8f50