Image Edge Detection Filter

WAS's own description is a single line - "Detect edges in a image" - and it's genuinely that plain: feed it an image, get back an edge map, usually bright lines on a dark background marking where contrast jumps sharply from one pixel to the next. It's a generic edge filter, not a named algorithm you'd cite by paper title.

How it's different from Canny

The pack also ships Image Canny Filter, which implements the Canny edge detector specifically - the algorithm that's become the de facto standard for ControlNet conditioning across the whole ecosystem, precisely because it's tunable (low/high threshold pair) and well understood. This node is the plainer sibling: a general edge-detection pass without Canny's specific multi-stage approach or its named threshold controls. If you're prepping a ControlNet input, Canny is what everyone actually uses and what most workflows expect - the KB's own read on ControlNet tooling names the dedicated ControlNet Auxiliary Preprocessors pack as "functionally required for any serious ControlNet workflow," and Canny is the standard edge preprocessor inside it. This node is better thought of as a quick, no-setup outline pass for previewing structure or building a stylized line-art look, not a ControlNet-grade preprocessor.

Why you'd reach for it anyway

Sometimes you don't need Canny's precision - you just want to see where the strong edges in an image are, fast, without picking two threshold values. A quick sanity check on composition, a lightweight line-art effect for a stylized output, or a cheap way to spot where detail concentrates in an image before deciding what to crop or upscale.

The inputs and outputs that matter

An image in, an edge-map image out. I'm working from the node's stated job rather than a confirmed field-by-field schema, so I won't invent specific threshold or sensitivity widget names here - the real controls (if any beyond the basic pass) are on the node itself once you drop it on the canvas.

How to install it

Ships with the full WAS Node Suite:

• ComfyUI Manager - search "WAS Node Suite", install, restart.
• Manual - cd ComfyUI/custom_nodes && git clone https://github.com/WASasquatch/was-node-suite-comfyui, install requirements.txt (portable: python_embeded\python.exe -s -m pip install -r requirements.txt), restart.

Pure image processing, no model download for this node.

Common issues & troubleshooting

It's noisier or less clean than you expected. Generic edge detection tends to pick up more incidental texture and noise than Canny's threshold-gated approach does - that's inherent to the simpler method, not a bug. If you need a clean, controllable edge map (especially for ControlNet), switch to Image Canny Filter or the ControlNet Aux pack's Canny preprocessor instead.

"Wrong" edges on a noisy or low-contrast source. Edge detection of any flavor is sensitive to the input's own noise and contrast. A flat, low-detail generation gives you a sparse, unsatisfying edge map - that's the source image, not this filter misbehaving.

Standard pack caveat: WAS Node Suite has been unmaintained since December 2023, and the recurring issue is the whole suite failing to import after a ComfyUI update. Reinstall requirements.txt against your ComfyUI Python if the suite won't load - this filter itself has no dependency of its own beyond the pack's core requirements.