SDPoseDrawKeypoints - ComfyUI Built-in Node Documentation

The SDPoseDrawKeypoints node takes pose estimation data (keypoints) and draws them as a visual skeleton on a blank canvas. It allows you to selectively draw different parts of the pose, such as the body, head, hands, face, and feet, with customizable line widths and point sizes. The resulting image can be used for visualization or as input for other nodes that require a pose image.

Inputs

Parameter	Description	Data Type	Required	Range
`keypoints`	The pose keypoint data to be drawn. This data typically comes from a pose detection node and can contain one or more frames.	POSE_KEYPOINT	Yes	-
`draw_body`	Controls whether the main body skeleton is drawn (default: True).	BOOLEAN	No	-
`draw_hands`	Controls whether the hand keypoints are drawn (default: True).	BOOLEAN	No	-
`draw_face`	Controls whether the face keypoints are drawn (default: True).	BOOLEAN	No	-
`draw_feet`	Controls whether the foot keypoints are drawn (default: False).	BOOLEAN	No	-
`stick_width`	The width of the lines used to draw the body and head skeleton (default: 4).	INT	No	1 to 10
`face_point_size`	The size of the points used to draw the face keypoints (default: 3).	INT	No	1 to 10
`score_threshold`	The minimum confidence score a keypoint must have to be drawn. Keypoints with scores below this value are ignored (default: 0.3).	FLOAT	No	0.0 to 1.0
`draw_head`	Controls whether the head keypoints (nose, eyes, ears) are drawn (default: True).	BOOLEAN	No	-

Note: If the keypoints input is empty or None, the node will output a blank 64x64 image. Note: draw_body and draw_head work independently. When draw_head is disabled, the head keypoints are not drawn even if draw_body is enabled. When draw_body is disabled but draw_head is enabled, only the head keypoints and the neck point are drawn. If both are disabled, no body or head keypoints are drawn.

Outputs

Output Name	Description	Data Type
`output`	An image with the drawn pose keypoints. The image dimensions match the `canvas_height` and `canvas_width` specified in the input keypoint data. When the input contains multiple frames, a batch of images is returned.	IMAGE

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): 2b2b9530b55c56e278666bd5d139bb6a1bb503b75b948a89266b9982b5a295e4

​Inputs

​Outputs

Inputs

Outputs