Nodes/ComfyUI Face Detection Node/Face Detection and Crop v2
ComfyUI Node

Face Detection and Crop v2

Face Detection v2 — Auto-Padding, Temporal Smoothing, Aspect Ratios, Batch Processing

By Limbicnation·Created about a year ago·Updated 4 months ago· 8
Face Detection and Crop v2
  • image
  • cropped_faces
  • face_metadata
detection_threshold0.80
min_face_size64
auto_padding_ratio35
aspect_ratioauto
output_modelargest_face
classifier_typedefault
temporal_smoothing0
output_height512
instance_id0
face_output_formatstrip
padding0
Categoryimage/processing

Inputs (12)

NameTypeDefaultDescription
imageIMAGE
detection_thresholdFLOAT0.800.1–1
min_face_sizeINT6432–512
auto_padding_ratioINT350–100Padding as % of detected face size
aspect_ratioCOMBOauto5 options: auto, 1:1, 9:16, 16:9, 4:3
output_modeCOMBOlargest_face2 options: largest_face, all_faces
classifier_typeCOMBOdefault2 options: default, alternative
temporal_smoothingoptINT00–1000=disabled (image mode) | 1-100: EMA smoothing strength
output_heightoptINT512256–2048
instance_idoptSTRING0Unique ID for temporal smoothing (share across frames in same video sequence). Use '0' for image mode.
face_output_formatoptCOMBOstripFormat for multiple faces (only applies with all_faces). Accepts 'strip' or 'individual'.
paddingoptINT00–256Legacy padding in pixels. If >0, overrides auto_padding_ratio.

Outputs (2)

NameTypeDescription
cropped_facesIMAGE
face_metadataFLOAT