ComfyUI Extension: ComfyUI_Swwan

Authored by aining2022

Created

Updated

2 stars

Collection of image processing utility nodes for ComfyUI including intelligent cropping, aspect ratio scaling, and crop box restoration capabilities. (Description by CC)

Custom Nodes (0)

    README

    ComfyUI LayerStyle Utility Nodes

    A collection of essential image processing utility nodes for ComfyUI, migrated from the popular ComfyUI_LayerStyle project. These nodes provide powerful image cropping, scaling, and restoration capabilities for advanced ComfyUI workflows.

    ✨ Features

    • 🎯 Smart Image Cropping: Intelligent mask-based cropping with multiple detection modes
    • 📐 Aspect Ratio Scaling: Flexible image scaling with aspect ratio preservation
    • 🔄 Crop Box Restoration: Seamlessly restore cropped images back to original canvas
    • 🎛️ Fast Groups Muter: Quick control of node groups mute/unmute state (from rgthree-comfy)
    • ⚡ Optimized Performance: Lightweight implementation with minimal dependencies
    • 🛠️ Workflow Integration: Designed for seamless integration in complex ComfyUI pipelines

    � Node List

    LayerUtility: CropByMask V2

    Intelligently crop images based on mask regions with advanced detection algorithms.

    Features:

    • Three detection modes: mask_area, min_bounding_rect, max_inscribed_rect
    • Customizable margin reserves (top, bottom, left, right)
    • Round dimensions to multiples (8, 16, 32, 64, 128, 256, 512)
    • Optional manual crop box input
    • Returns cropped image, mask, crop box coordinates, and preview

    Use Cases:

    • Extract masked regions for focused processing
    • Prepare images for inpainting workflows
    • Optimize processing area to reduce computation

    LayerUtility: RestoreCropBox

    Restore cropped images back to their original canvas position.

    Features:

    • Paste cropped images back to original coordinates
    • Support for mask-based compositing
    • Automatic alpha channel handling
    • Batch processing support
    • Mask inversion option

    Use Cases:

    • Restore processed regions to original image
    • Complete crop → process → restore workflows
    • Seamless image compositing

    LayerUtility: ImageScaleByAspectRatio V2

    Scale images to specific aspect ratios with multiple fitting modes.

    Features:

    • Preset aspect ratios: 1:1, 3:2, 4:3, 16:9, 21:9, 3:4, 9:16, and more
    • Custom aspect ratio support
    • Three scaling modes: letterbox, crop, fill
    • Scale to specific side (longest, shortest, width, height)
    • Round dimensions to multiples
    • SSAA (Super-Sampling Anti-Aliasing) support

    Use Cases:

    • Prepare images for specific output formats
    • Maintain aspect ratios during processing
    • Create consistent image dimensions for batch processing

    Fast Groups Muter (rgthree)

    Quick control of workflow groups' mute/unmute state. Virtual node for UI control only.

    Features:

    • Auto-detect all groups in workflow
    • Toggle mute state for all nodes in a group
    • Filter groups by color or title (regex support)
    • Multiple sort options (position, alphabetic, custom)
    • Batch operations (mute all, enable all, toggle all)
    • Quick navigation to group location

    Use Cases:

    • Quickly enable/disable entire sections of workflow
    • Test different workflow branches
    • Organize complex workflows with groups
    • Debug by isolating specific groups

    Note: This is a virtual node (frontend only) ported from rgthree-comfy. See FAST_GROUPS_MUTER_README.md for detailed usage.

    Fast Muter (rgthree)

    Quick control of connected nodes' mute/unmute state. Virtual node for UI control.

    Features:

    • Auto-detect connected nodes
    • Toggle mute state for each connected node
    • Batch operations (mute all, enable all, toggle all)
    • Toggle restrictions (default, max one, always one)

    Use Cases:

    • Control multiple nodes from a single point
    • Test different processing paths
    • Quickly enable/disable node chains

    Image Comparer (rgthree)

    Compare two images side-by-side with interactive slider or click mode.

    Features:

    • Slide mode: hover to compare images
    • Click mode: click to switch between images
    • Support for image batches
    • Automatic image selection from batches

    Use Cases:

    • Compare before/after processing results
    • Evaluate different model outputs
    • Quality control and A/B testing

    Seed (rgthree)

    Enhanced seed node with special functions for randomization and control.

    Features:

    • Random seed generation
    • Increment/decrement seed values
    • Fixed random seed option
    • Last seed tracking and reuse
    • Server-side random generation fallback

    Use Cases:

    • Consistent reproducible results
    • Systematic seed exploration
    • Quick randomization for testing
    • Seed value management

    🚀 Installation

    Method 1: ComfyUI Manager (Recommended)

    1. Open ComfyUI Manager
    2. Search for "LayerStyle Utility"
    3. Click Install
    4. Restart ComfyUI

    Method 2: Manual Installation

    # Navigate to ComfyUI custom_nodes directory
cd ComfyUI/custom_nodes

# Clone this repository
git clone https://github.com/YOUR_USERNAME/ComfyUI_LayerStyle_Utility

# Install dependencies
cd ComfyUI_LayerStyle_Utility
pip install -r requirements.txt

# Restart ComfyUI

    📦 Dependencies

    • torch - PyTorch for tensor operations
    • torchvision - Computer vision utilities
    • Pillow - Image processing library
    • numpy - Numerical computing
    • opencv-python - Advanced image processing

    All dependencies are automatically installed via requirements.txt.

    � Usage Examples

    Example 1: Crop → Process → Restore Workflow

    [Load Image] → [CropByMask V2] → [Your Processing Node] → [RestoreCropBox] → [Save Image]
                      ↓
                  [Load Mask]

    This workflow allows you to:

    1. Crop a specific region using a mask
    2. Process only the cropped area (faster, more efficient)
    3. Restore the processed region back to the original image

    Example 2: Aspect Ratio Standardization

    [Load Image] → [ImageScaleByAspectRatio V2] → [Your Model] → [Save Image]

    Perfect for:

    • Preparing images for models that require specific dimensions
    • Creating consistent output sizes
    • Maintaining aspect ratios during batch processing

    Example 3: Advanced Inpainting Pipeline

    [Load Image] ──┬─→ [CropByMask V2] → [Inpainting Model] → [RestoreCropBox] ──→ [Save Image]
               │                                                    ↑
[Load Mask] ───┴────────────────────────────────────────────────────┘

    🎯 Node Parameters

    CropByMask V2

    • image: Input image tensor
    • mask: Mask defining the crop region
    • invert_mask: Invert the mask (default: False)
    • detect: Detection mode (mask_area, min_bounding_rect, max_inscribed_rect)
    • top/bottom/left/right_reserve: Margin pixels to add around detected region
    • round_to_multiple: Round dimensions to specified multiple
    • crop_box (optional): Manual crop box coordinates

    RestoreCropBox

    • background_image: Original full-size image
    • croped_image: Cropped image to restore
    • crop_box: Crop box coordinates from CropByMask V2
    • croped_mask (optional): Mask for compositing
    • invert_mask: Invert the mask (default: False)

    ImageScaleByAspectRatio V2

    • aspect_ratio: Target aspect ratio (original, custom, or preset)
    • proportional_width/height: Custom aspect ratio values
    • fit: Scaling mode (letterbox, crop, fill)
    • scale_to_side: Which side to scale to (longest, shortest, width, height)
    • scale_to_length: Target length in pixels
    • round_to_multiple: Round dimensions to specified multiple
    • image/mask: Input image or mask tensor

    🛠️ Technical Details

    Detection Modes Explained

    • mask_area: Uses the entire mask area as crop region
    • min_bounding_rect: Finds the minimum bounding rectangle around mask
    • max_inscribed_rect: Finds the largest rectangle that fits inside mask

    Scaling Modes Explained

    • letterbox: Fit image within target size, add padding if needed
    • crop: Fill target size, crop excess if needed
    • fill: Stretch image to exactly fill target size

    🤝 Credits

    These nodes are migrated from the excellent ComfyUI_LayerStyle project by chflame163. We've extracted and optimized these specific utilities for users who need these functions without the full LayerStyle suite.

    Original project: https://github.com/chflame163/ComfyUI_LayerStyle

    📄 License

    This project maintains the same license as the original ComfyUI_LayerStyle project.

    🐛 Issues & Support

    If you encounter any issues or have questions:

    1. Check the Issues page
    2. Create a new issue with detailed description
    3. Include your ComfyUI version and error logs

    🌟 Contributing

    Contributions are welcome! Please feel free to:

    • Report bugs
    • Suggest new features
    • Submit pull requests
    • Improve documentation

    📝 Changelog

    v1.0.0 (Initial Release)

    • Migrated CropByMask V2 node
    • Migrated RestoreCropBox node
    • Migrated ImageScaleByAspectRatio V2 node
    • Created standalone utility module
    • Optimized dependencies

    Note: This is a focused utility package. For the complete LayerStyle suite with 100+ nodes, please visit the original ComfyUI_LayerStyle project.