ComfyUI-LoadImageEnhanced

ComfyUI LoadImageEnhanced

This custom node extends ComfyUI's image loading functionality with filename output and original filename tracking.

Features

LoadImageEnhanced

• Enhanced Load Image Node: Based on the original ComfyUI LoadImage node, inheriting from ComfyNodeABC
• Compatible: Maintains all original functionality (IMAGE and MASK outputs)
• Filename Output: Returns the filename of the loaded image as a STRING output
• Original Filename Tracking: Preserves the original filename even when the image is modified by tools like MaskEditor
• Multi-frame Support: Supports animated images (e.g., GIFs) with multi-frame iteration
• EXIF Handling: Automatically applies EXIF orientation rotation
• Subfolder Listing: Recursively lists images from subfolders (excludes clipspace, 3d, pasted, and any folder/file starting with .)

Installation

1. Place the entire ComfyUI-LoadImageWithFilename folder in your ComfyUI custom_nodes directory
2. Restart ComfyUI
3. The new node will appear in the "image" category

Usage

LoadImageEnhanced

• Input: Select an image file from the dropdown (includes images from subfolders)
• Inputs:
  • image — Image filename to load
  • original_filename — Text field that stores the original selected filename (editable by user)
• Outputs:
  • image: The loaded image tensor
  • mask: The image mask (if available)
  • filename: The filename of the loaded image

Requirements

• ComfyUI
• PIL (Pillow)
• PyTorch
• NumPy

Notes

Filename Tracking

The filename output uses the value of the original_filename widget. This widget is automatically updated when you select a new image from the dropdown, but is not changed by downstream tools like MaskEditor. This means:

• When you first select an image (e.g., characters/hero.png), the filename output is hero.png
• If MaskEditor saves the edited image to clipspace-painted-xxx.png, the dropdown updates but the original_filename widget stays as hero.png
• The filename output continues to return hero.png even after mask edits

Frontend Extension (Automatic)

The node includes a frontend extension that automatically syncs the original_filename widget when you change the dropdown. It's bundled with the custom node and loads automatically via WEB_DIRECTORY.

Without the extension, the original_filename field is still available but must be manually edited.

• The node recursively lists all images in the input folder and its subfolders
• Images inside clipspace, 3d, pasted, and any hidden (starting with .) folders are excluded from the dropdown listing
• Any file starting with . (e.g., .DS_Store) is also excluded from the listing
• However, if a downstream node (e.g., mask editor) passes an image path from an excluded folder, the node will still load it correctly
• Images with multiple frames (e.g., animated GIFs) are processed frame-by-frame and concatenated
• If no alpha channel is found in an image, a default 64x64 mask is returned
• The node automatically handles EXIF orientation for proper image display
• Error messages are printed to console for any files that fail to load

Based On

• ComfyUI LoadImage Node