Skip to main content
The DrawBBoxes node visualizes object detection results by drawing bounding boxes, labels, and confidence scores onto an image. If no input image is provided, it creates a blank canvas large enough to contain all the drawn boxes. It supports batch processing, allowing you to draw different sets of detections for multiple images or repeat the same detections across a batch.

​
Inputs

ParameterDescriptionData TypeRequiredRange
imageThe input image(s) to draw the bounding boxes onto. If not provided, a blank canvas will be generated.IMAGENo-
bboxesA list of bounding box dictionaries. Each dictionary should contain keys for x, y, width, height, and optionally label and score.BOUNDINGBOXYes-
Input Constraints:
  • The bboxes input is required and must be provided.
  • The node automatically handles different input formats for bboxes. A single dictionary will be applied to all images in the batch. A flat list of dictionaries will be treated as the same set of detections for every image. A list of lists allows you to specify different detections for each image in the batch.
  • If an image is not provided, the node will create a blank image with dimensions large enough to fit all provided bounding boxes, with a default minimum size of 640x640.
  • The label field in each bounding box dictionary is matched against the COCO dataset classes. If the label is not a recognized COCO class, it will be ignored and no text label will be drawn for that box.

​
Outputs

Output NameDescriptionData Type
out_imageThe output image(s) with the drawn bounding boxes, labels, and confidence scores overlaid.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): ce5da10ec83e579515d70b7184fc78dfa04df4c514faccdcb75130212f56c84f