Save 4 hours! We auto-setup your workflow! Free!

Drop your workflow.json — we handle every dependency, custom node, and model. Just open the link and run.

Auto-Setup Workflow Json (Free) Now!
ComfyUI > Nodes > comfyui-image-process

ComfyUI Extension: comfyui-image-process

Repo Name

comfyui-image-process

Author
wakaura-asaho (Account age: 544 days)
Nodes
View all nodes(1)
Latest Updated
2026-09-27
Github Stars
0.0K

How to Install comfyui-image-process

Install this extension via the ComfyUI Manager by searching for comfyui-image-process
  • 1. Click the Manager button in the main menu
  • 2. Select Custom Nodes Manager button
  • 3. Enter comfyui-image-process in the search bar
After installation, click the Restart button to restart ComfyUI. Then, manually refresh your browser to clear the cache and access the updated list of nodes.

Visit ComfyUI Online for ready-to-use ComfyUI environment

  • Free trial available
  • 16GB VRAM to 80GB VRAM GPU machines
  • 400+ preloaded models/nodes
  • Freedom to upload custom models/nodes
  • 200+ ready-to-run workflows
  • 100% private workspace with up to 200GB storage
  • Dedicated Support

Run ComfyUI Online

comfyui-image-process Description

comfyui-image-process adds ComfyUI nodes to correct color casts, smooth and unify color patches, manage transparency and ICC profiles, and export images in multiple formats or multi-size ICO files.

comfyui-image-process Introduction

comfyui-image-process is a collection of ComfyUI nodes for refining AI-generated images and saving them in formats suited to different uses. You can use it to clean up unwanted color in nearly gray areas, simplify patches of color, preserve transparency, embed a color profile, or prepare a Windows icon with several sizes in one file.

The nodes fit into a visual workflow: connect an image, adjust the settings you need, and pass the result to another node or a saver. You do not need to use every feature. For example, you might use Achromatic Stabilizer to remove a faint color cast from a white background, or use ICO Packer only when you need an app icon.

Example image-processing workflow

How comfyui-image-process Works

Think of the extension as a set of finishing tools for your image:

  1. Choose what needs attention. A processing node receives an image and examines its pixels. Some nodes can also use a mask to limit where they work.
  2. Control how much changes. Settings such as a saturation threshold, color tolerance, or smoothing size determine which pixels are affected and how strongly.
  3. Keep or prepare transparency. Where supported, a mask can help preserve an existing transparent area or supply an alpha channel when saving. An alpha channel tells an image viewer which parts of an image are transparent.
  4. Save for the intended use. Choose a file format and its available options, such as PNG compression, JPEG quality, TIFF compression, or multiple icon sizes.

For instance, an AI-generated illustration might have a slightly green area that should be neutral gray. Achromatic Stabilizer identifies pixels with very little color and makes the affected pixels neutral. You can then send the result to a save node. If the illustration needs a transparent background, connect the appropriate mask when using the relevant alpha options.

comfyui-image-process Features

Achromatic Stabilizer

Achromatic Stabilizer corrects color instability in areas that should look nearly gray, white, or black. It detects pixels with very low saturation and normalizes affected pixels to neutral gray.

  • Saturation Threshold controls which low-saturation pixels are treated as artifacts. The documented starting point is 0.08. The author suggests trying 0.05 for more aggressive correction or 0.15 for a more conservative result.
  • Smooth Transitions blends corrected areas with their surroundings. Smoothing Kernel Size controls the blending area: 3–5 is suggested for subtle changes, while 7–15 gives stronger smoothing but may blur regions.
  • Alpha options let you preserve transparency or invert the alpha channel when needed.
  • Dual output provides the processed image and an alpha mask separately, so you can route them to later nodes.

This node is useful when a pale background has faint colored noise or when dark, nearly colorless areas look inconsistent. Compare the result with your original image after changing the threshold, especially if the artwork contains intentional subtle color.

Color Patch Flatten

Color Patch Flatten makes similar colors within a patch more consistent. It works with three parts of color, which you can enable independently:

  • Hue (H): the color family, such as red or blue.
  • Saturation (S): how vivid or muted the color is.
  • Value (V): its brightness.

You can flatten only the aspects that need work. For example, adjusting saturation without adjusting hue can make a painted region more even without intentionally changing its color family. Tolerance controls which values qualify; according to the author, larger values produce a weaker flattening effect. An alpha mask can restrict both the color calculation and the changes to a selected region.

Color Patch Merge

Color Patch Merge groups nearby, similar colors into more unified areas. It offers two approaches:

ApproachWhat it doesWhen to try it
SmoothUses edge-preserving smoothing, controlled by Neighborhood.When you want gentler color transitions without simply blurring every edge.
UnifyBuilds more solid color regions, controlled by Iterations and Minimal Area.When you want flatter patches while protecting strong edges and small regions below the minimum-area setting.

The node can optionally use Lab color space, a way of representing color that can help group colors according to perceived similarity. If an effect is too strong, compare the two approaches and adjust their respective controls.

Load ICC Profile

An ICC profile describes how colors should be interpreted across devices or applications. Load ICC Profile lets you select a valid .icc or .icm file from models/icc_profiles and pass its data to Save Image Advanced for embedding. It also outputs profile information, including the model, manufacturer, description, and copyright fields.

This node loads a color-profile file; the extension does not list separate image-generation models. BMP and TGA do not support ICC profile embedding in this extension.

ICC profile save workflow

Save Image Advanced

Save Image Advanced gives you more control over your final file than a basic save step. It supports JPG, PNG, WebP, TIFF, BMP, and TGA, with options that vary by format:

  • Transparency: join an image with a mask as an alpha channel, or invert that alpha channel. BMP offers 24-bit and 32-bit output; the 32-bit option supports alpha.
  • Color profile: embed a profile supplied by Load ICC Profile where the format supports it.
  • Metadata: Disable Metadata removes workflow metadata to reduce file size, but also means that metadata will not be available for reading the workflow later.
  • File settings: adjust PNG compression from 0–9, JPEG/WebP quality from 1–100, DPI metadata from 1–600 (default 300), TIFF compression, or TGA RLE compression.
  • Reuse: Save to Input Folder copies the saved file to ComfyUI’s input folder.

Filename Prefix can include tokens that change with the image or save time. Right-click the node and choose Insert Filename Token to add one without typing it manually. These tokens are provided by the extension independently of ComfyUI’s core filename behavior.

TokenWhat it addsExample result
%date:FORMAT%A date or time in your chosen format%date:yyyy-MM-dd% → 2026-06-28
%year%, %month%, %day%Year, month, or day2026, 06, 28
%hour%, %minute%, %second%Time components21, 05, 00
%width%, %height%Image dimensions in pixels1024, 1024

For a custom date, %date:FORMAT% accepts yyyy, yy, MM, dd, HH, hh, mm, and ss. For example, renders/%date:yyyy-MM-dd%/shot organizes saves under a dated path.

Filename token menu

Advanced image save workflow

Note: The extension’s Save Image (Advanced) node has the internal class name SaveImageAdvancedCustom. ComfyUI core also has a node named Save Image Advanced, so check that you have selected the intended node when following an example workflow.

Dedicated Format Savers

The extension also provides format-specific savers in simple and advanced variants. Choose a simple saver for a straightforward export, or an advanced variant when you need its additional settings.

FormatAvailable purpose and controls
JPGQuick saving, or advanced EXIF and DPI control.
BMPStandard saving, or 32-bit alpha and DPI support.
TIFFStandard saving, or advanced compression and EXIF options.
TGAStandard saving, or RLE compression and alpha support.
AVIFStandard saving, or advanced quality and speed control.

For TIFF, the author recommends tiff_ccitt for black-and-white documents or line art. For photographs, try tiff_lzw, tiff_deflate, or tiff_adobe_deflate.

Simple save workflow

ICO Packer and ICO Savers

A Windows .ico file can contain an icon at several resolutions. This helps the icon display at different sizes without relying on a single image being resized for every situation.

ICO Packer prepares the images for Save Image Advanced (ICO):

  • Add image_<size> and optional mask_<size> inputs from the node’s context menu. The supported standard sizes are 16, 24, 32, 48, 64, 128, and 256 px, with up to seven unique sizes.
  • Use Add Input for one size, Batch Add Input to queue several sizes, or Remove Input to remove a slot.
  • Enable Crop Images to crop an oversized frame to its slot size when both frame dimensions are at least that size. A connected mask uses the same crop window.
  • Set Reference Point to center for a centered crop or top-left when alignment with the top-left corner matters. This control applies only when cropping is enabled.
  • Connect the packer’s single ICO_PACKER output to Save Image Advanced (ICO).

The two ICO savers serve different workflows:

  • Save Image (ICO) accepts a batch of images. It bundles valid square frames in the supported sizes into one .ico file and can optionally save invalid frames as separate _INVALID.png files.
  • Save Image Advanced (ICO) uses ICO Packer. It offers 8-bit (256-color palette), 16-bit (RGB), or 32-bit (RGBA) color depth; uncompressed BMP frames or ZIP PNG-compressed frames for Windows Vista and later; per-size mask and alpha-inversion options; optional sorting from smallest to largest; copying to the input folder; and optional PNG saves for size-mismatched images.

A typical advanced workflow is resize or crop each source → add the needed sizes to ICO Packer → connect it to Save Image Advanced (ICO).

ICO save workflow

What's New with comfyui-image-process

Version 1.3.0

The supplied release notes describe these changes:

  • ICO Packer was added. You can collect images and optional masks by icon size, manage slots from the node’s context menu, and send the collection through one output. This makes multi-size icon workflows easier to organize.
  • Save Image Advanced (ICO) now takes the ICO Packer connection. Its former per-size image and mask inputs were removed. Existing workflows using those inputs need to connect their sources to ICO Packer instead.
  • Crop Images and Reference Point were added to ICO Packer. Oversized frames and their masks can be cropped for each slot, using either a centered or top-left crop. This can reduce the amount of resizing or cropping you need to arrange elsewhere in an icon workflow.

No earlier version history is included in the supplied documentation.

Troubleshooting comfyui-image-process

My ICO workflow no longer connects to Save Image Advanced (ICO)

Version 1.3.0 removed the saver’s separate per-size inputs.

  1. Add an ICO Packer node.
  2. Use Add Input or Batch Add Input to create the sizes you need.
  3. Connect each image, and any corresponding mask, to its size slot.
  4. Connect the packer’s ICO_PACKER output to the saver’s ICO Packer input.

The ICO example workflow and the ICO workflow screenshot show the intended arrangement.

My icon image is rejected or saved as an invalid PNG

ICO inputs are checked against the supported square dimensions: 16, 24, 32, 48, 64, 128, or 256 px.

  1. Check the width and height of the image assigned to the slot.
  2. Resize or prepare it for the slot’s dimensions.
  3. If the source is larger in both dimensions, try Crop Images on ICO Packer and choose the appropriate Reference Point.
  4. Run the workflow again. If Save Invalid as PNG is enabled, inspect the fallback PNG to identify the frame that did not match.

I cannot change Reference Point

Reference Point is disabled when Crop Images is off. Enable Crop Images on ICO Packer, then choose center or top-left.

My TIFF or TGA preview does not look right in ComfyUI

ComfyUI may display a PNG preview copy rather than previewing the saved TIFF or TGA directly. Check the actual saved file in an application that supports your chosen format before judging the export.

My image saved without its intended transparency

  1. Check that the relevant image and mask are connected.
  2. Check Join Alpha Channel and Invert Alpha settings on the saver you are using.
  3. For BMP, choose 32-bit output when you need alpha; the documented 24-bit option does not provide it.
  4. For an ICO, check the optional mask attached to each applicable size in ICO Packer.

Frequently Asked Questions

Do I need every node to improve an image?
No. Use only the nodes that address your image. Color correction, color-patch processing, profile loading, and saving can each be used for different purposes.

Which color-patch node should I try first?
Try Color Patch Flatten when you want to even out selected hue, saturation, or brightness values within patches. Try Color Patch Merge when you want neighboring similar colors to form smoother or more solid regions.

Why did my saved file get smaller after I enabled Disable Metadata?
That option discards workflow metadata. The tradeoff is that the discarded metadata will not be available for later workflow inspection.

Can I embed an ICC profile in BMP or TGA?
No. The author notes that those formats do not support ICC profile embedding in this extension.

Does the extension include image-generation models?
No separate image-generation models are identified in the supplied documentation. Load ICC Profile selects color-profile files, not AI image models.

Learn More about comfyui-image-process

The author describes the extension as designed for modern ComfyUI V3 nodes and reports testing with frontend version 1.37.11 or later and base version 0.12.3 or later.

comfyui-image-process Related Nodes

RunComfy
Copyright 2025 RunComfy. All Rights Reserved.

RunComfy is the premier ComfyUI platform, offering ComfyUI online environment and services, along with ComfyUI workflows featuring stunning visuals. RunComfy also provides AI Models, enabling artists to harness the latest AI tools to create incredible art.

comfyui-image-process detailed guide | ComfyUI