Face Detection

Overview

Face Detection node locates faces in an input image and returns bounding boxes, optional facial landmarks, and an overlay image.

Use this node to feed detected face regions into downstream nodes such as Face Recognition, blur/mosaic filters, or zone-based logic.

Input

Input Image

image required

The image frame to analyze. Connect this to an image output from an upstream camera or processing node.

Detection Model

string required

Face detection backend to use.

Values:

• ONNX_YUNET (default) — faster, recommended for most use cases.
• TFDNN — TensorFlow DNN-based detector; use when ONNX_YUNET does not produce satisfactory results on your hardware.

Confidence Threshold

number

Minimum confidence score to keep a detected face. See Confidence Threshold for tuning guidance.

Default: 0.9

NMS Threshold

number

Non-Maximum Suppression threshold for overlapping face boxes. See NMS Threshold for tuning guidance.

Default: 0.3

YuNet Top K

integer

Applies only when Detection Model is ONNX_YUNET. Sets the maximum number of candidate face boxes to keep before NMS is applied. Lower values reduce computation at the cost of potentially missing distant or small faces.

Default: 5000

Overlay Results

boolean required

Whether to draw face bounding boxes on the output frame. See Overlay Results.

Draw Landmarks

boolean required

When enabled, draws the five facial landmark points (eye corners, nose tip, and mouth corners) on the overlay for each detected face.

Output

Overlay Image

image

Output frame from the node. If overlays are enabled, face bounding boxes and optional landmarks are drawn on this frame.

Detected Count

integer

Number of faces detected in the current frame.

Detected Faces

array

Array of face detection objects. Each object contains:

• bbox array: Bounding box [x, y, width, height] in image pixel coordinates.
• landmarks array: Five landmark points [[x, y], ...] when Draw Landmarks is enabled.
• confidence number: Detection confidence score.