🌟 Neon Glow

Neon Glow finds the edges in your image and lights them up like a neon sign - exactly the treatment that sells cyberpunk cityscapes, sci-fi character portraits, and anything you want to look backlit in electric color. Unlike a generic bloom, this is edge-aware: it detects where your subject stops and the background starts, then wraps a glow around that boundary. That's what separates it from just cranking up brightness.

It's one of the Light Effects in the pack, and it sits at the finishing end of your graph - glow effects always go after compositing, before Save.

How it works

The pipeline is honest about being a classic trick. It converts the image to grayscale and runs a Canny edge detector (the edge_threshold input sets the low and high hysteresis bounds). Those edges become a mask, which is then painted in your chosen glow_color and blurred with a Gaussian - a small blur for the inner glow, a larger one for the outer glow, so you get that two-layer neon halo. The blurred layers are added back over the original scaled by intensity.

Two details worth knowing. First, glow_size controls the halo radius - bump it up for a heavy, blown-out neon and down for a crisp tight line. Second, if pulsate is on, the brightness breathes using real time, which is a nice touch for animated frames and a mild curiosity on a still image.

The inputs that matter

• glow_color - eight presets: cyan, magenta, yellow, red, green, blue, purple, orange. Cyan is the default for a reason.
• intensity (0–2, default 0.8) - brightness of the glow. Over 1.5 starts to blow out the edges.
• glow_size (0.1–1, default 0.3) - halo thickness.
• edge_threshold (0.1–1, default 0.3) - sensitivity of the edge detection. Lower detects more (and noisier) edges; raise it if your background is busy and the glow looks messy.
• inner_glow / outer_glow (both on) - toggle the two halo layers independently. Inner-only gives a tight rim-light; outer-only gives a soft ambient spread.

Output is the glowing IMAGE.

Installing it

From ComfyUI-Image-Effects (Orion4D). ComfyUI Manager → search "Image Effects", or:

cd ComfyUI/custom_nodes
git clone https://github.com/orion4d/ComfyUI-Image-Effects
cd ComfyUI-Image-Effects
pip install -r requirements.txt

Restart, then find it under Add Node → Image Effects. No downloads beyond the usual NumPy/OpenCV/Pillow/SciPy stack.

Gotchas

The pack standard applies: only the first image of a batch is processed, and it's CPU NumPy, so on 4K images the double Gaussian blur is the slow bit. The real failure mode to watch for: Canny finds every edge, including noise and texture in the background. If your image is busy, raise edge_threshold toward 0.5–0.6 before you touch any other slider - a glowing background texture reads as a mistake instantly, while a clean rim on the subject looks intentional.