ComfyUI Deepface

ComfyUI nodes wrapping the deepface library.

Requirements

Install dependencies:

pip install -r requirements.txt

Nodes

Deepface Extract Faces

TODO: document

Deepface Verify

Given a set of input images and a set of reference (face) images, only output the input images with an average distance to the faces in the reference images less than or equal to the specified threshold. Output images are sorted by average distance to the reference image faces (nearest first).

Both inputs accept image batches.

• Outputs:
  • verified_images images that contain faces below the threshold, sorted from smallest distance to largest
  • verified_image_distances numbers of the average distance from the input image face to each reference image face
  • verified_image_verified_ratios the ratio of reference images that deepface marks as "verified" to the total number of reference images. As far as I can tell, deepface sets "verified" to true for distances under 0.3, but this might vary per detector.
  • rejected_images images that contain faces above (or equal to) the threshold value for distance, sorted from smallest distance to largest
  • rejected_image_distances numbers of the average distance from the input image face to each reference image face
  • rejected_image_verified_ratios the ratio of reference images that deepface marks as "verified" to the total number of reference images. As far as I can tell, deepface sets "verified" to true for distances under 0.3, but this might vary per detector.

Nodes from was-node-suite-comfyui and pythongosssss can be used to display the output distances:

Credits

Inspired by CeFurkan's use of deepface to evaluate finetuning outputs. 🙌