Processing

Canny Edge Detection Detector

Inputs

• Image(s): Input.
• Mask(s): (Optional) Mask.

Widgets

• Lower Threshold: Sensitivity Start.
• Upper Threshold: Sensitivity End.
• Dilate: Thicken edges.
• Kernel Size: Blur/Detail.
• Iterations: Dilation steps.
• Color Map: Visualization colors.

Outputs

• Image(s): Edge map.

Description

Classic computer vision algorithm that finds outline edges in an image.

Parameters

Widget	Type	Description
lowerThreshold	Int	Minimum intensity gradient to be considered an edge.
upperThreshold	Int	Maximum intensity gradient. Values between Lower and Upper are accepted if connected to a strong edge.
dilate	Bool	Thicken the detected edges.
kernelSize	Int	Blur kernel size to reduce noise before detection.
iterations	Int	Number of times to dilate (thicken) the edges.
colorMap	Dropdown	Color scheme for the output (e.g., ‘winter’, ‘bone’, ‘jet’).

---

Centroid Tracker Tracker

Inputs

• BBoxes: Detections per frame.

Widgets

• Max Distance: Radius for matching.
• Max Disappeared: Tolerance for lost tracking.
• Jitter Red. (min_cutoff): Smoothness.
• Lag Red. (beta): Responsiveness.

Outputs

• Filtered BBoxes: Result list.
• Total: Count of objects.

Description

Assigns persistent IDs to objects across video frames. If Box A in frame 1 is close to Box B in frame 2, they obtain the same ID.

---

Contour Extraction Detector

Inputs

• Image: Binary mask/image.

Widgets

• Min Area: Noise filter.

Outputs

• BBoxes: Bounding boxes of shapes.
• Contours: Geometric path data.

Description

Finds the shapes (contours) inside a binary image. Useful for turning a mask into vector data or bounding boxes.

---

Displacement Effect

Inputs

• Image(s): Base image.
• Displacement Map: Displacement Map.

Widgets

• Amount: Pixel distance.
• Angle: Direction (0-360).

Outputs

• Image: Distorted result.

Description

Uses the brightness of the Displacement Map to push pixels of the Image(s).

• Black (0) = No movement.
• White (1) = Maximum displacement (defined by Scale).
• Direction is controlled by the Angle widget.

---

DW Pose Detector

Inputs

• image_batch: Source images.

Widgets

• Min Confidence: Accuracy filter.
• Draw Skeleton: Overlay visualization.
• Detect Hand: Enable hand keypoints.
• Detect Face: Enable face landmarks.

Outputs

• images: The colorful “skeleton” preview.
• json_data: Raw keypoints.

Description

Accurately detects human poses (Stickman representation). Essential for ControlNet OpenPose workflows.

Parameters

Widget	Type	Default	Description
Min Confidence	Float	0.3	Minimum score to consider a person detected.
Draw Skeleton	Bool	True	Draws the colorful stickman overlay on the output image.
Detect Hand	Bool	True	detects fingers and hands (slower).
Detect Face	Bool	True	detects facial landmarks (slower).

---

Grid Sampler Sampler

Inputs

• Image(s): Source image.
• Tiles: Regions to sample.

Widgets

• Sample Method: Luminosity/Sat/Alpha.
• Monochromatic: Black & White.

Outputs

• Images: Sampled patches.
• Colours: Average RGB Colors.
• Values: Number values.

Description

Samples the average color or value of specific regions (Tiles) from the image.

---

MOG2 Background Subtraction Detector

Inputs

• Image: Video frame.

Widgets

• History: frames to remember.
• Threshold: Sensitivity.
• Detect Shadows: Enable shadow detection.

Outputs

• Mask: Moving Object Mask.

Description

Automatically “learns” the static background of a video and outputs a mask of anything that is moving (Foreground).

---

Motion Mask Filter

Inputs

• Image: Current frame.

Widgets

• Threshold: Difference sensitivity.

Outputs

• Mask: Motion area.

Description

Simple “Frame Difference” motion detection. Subtracts Frame B from Frame A to see what changed.

---

Proximity Linker Generator

Inputs

• centers: A list of coordinates.

Widgets

• Threshold: Link radius.
• Max Connections: Limit density.

Outputs

• lines: Network of lines.

Description

The “Plexus” effect. Draws lines between any two points that are close enough to each other. Great for visualizing tracking data.

Parameters

Widget	Type	Description
Threshold	Float	Maximum distance between two points to create a connection.
Max Connections	Int	Maximum number of lines per point.

---

Stop Motion Node Effect

Inputs

• Image: Input video or image sequence.

Widgets

• Frame Rate: Target FPS (e.g., 12 for stop motion feel).
• Jitter Amount: Random pixel shift for “hand-made” look.
• Mode: Stutter (skip frames) or Normal.

Outputs

• Image: Processed stop motion sequence.

Description

Simulates a stop-motion animation effect by reducing the frame rate and optionally adding jitter to mimic hand-placed imperfections.

Parameters

Widget	Type	Description
Frame Rate	Int	The target frame rate for the stutter effect (e.g., 12fps is common for stop motion).
Jitter Amount	Float	Amount of random position shift applied to each frame step.
Mode	Dropdown	Stutter: Holds frames to match the target FPS.