ComfyUI Extension: ComfyUI-Image-Segmenting-Loader

Authored by GeekyGhost

Created

Updated

0 stars

Lets you send a section of an image to qwen edit or kontext to help isolate areas in need of change

Custom Nodes (0)

    README

    šŸŽ¬ Studio42 Image, Audio, and Video Editing Suite for ComfyUI

    This is a work in progress and not recommended for use at this time.

    Professional-grade custom nodes for advanced image and video editing workflows in ComfyUI

    Version ComfyUI License

    šŸŒŸ Overview

    Studio42 is a comprehensive suite of advanced custom nodes that brings professional-grade image and video editing capabilities to ComfyUI. Designed for efficiency, quality, and creative flexibility, this suite provides cutting-edge background removal, layer composition, and patch manipulation tools used in modern VFX and content creation workflows.

    šŸŽÆ Key Features

    šŸŽØ Professional Background Removal

    • Multi-method AI processing with BiRefNet, U2Net, and ISNet models
    • Industry-standard keying with professional chroma key, luma key, and difference key
    • Advanced spill suppression with luminance preservation
    • Multi-stage processing for complex scenes and challenging backgrounds
    • Batch processing with automatic memory optimization

    šŸŽ­ Advanced Layer Composition

    • Precise positioning with pixel-perfect coordinate control
    • Professional transformations including scaling, rotation, and anchor points
    • Comprehensive blend modes (Normal, Multiply, Screen, Overlay, Soft Light, etc.)
    • Edge processing with feathering and anti-aliasing
    • Real-time preview with visual feedback

    šŸŽÆ Interactive Patch Processing

    • Visual patch selection with grid overlay and shape tools
    • Multi-format support for images and video sequences
    • Shape-aware extraction (Rectangle, Circle, Rounded Rectangle)
    • Intelligent placement with automatic and manual positioning
    • Temporal consistency for video workflows

    šŸ“¦ Node Catalog

    1. šŸŽ¬ Studio42 Background Remover

    Professional multi-method background removal with VFX-quality results

    AI Methods:

    • birefnet-general - Best general purpose (recommended)
    • birefnet-portrait - Optimized for human portraits
    • birefnet-massive - Highest accuracy, largest dataset
    • birefnet-general-lite - Lightweight version for faster processing
    • u2net - Classic reliable model
    • u2net_human_seg - Human segmentation specialist
    • isnet-general-use - ISNet general model

    Professional Traditional Methods:

    • chroma_key_professional - Multi-stage chroma keying with spill suppression
    • luma_key_professional - Advanced luminance-based removal
    • difference_key - Background difference detection
    • hybrid_ai_traditional - AI + traditional combined approach

    Advanced Features:

    • Professional spill suppression with luminance preservation
    • Multi-stage processing pipeline (After Effects/Nuke quality)
    • Edge detection and refinement
    • Core matte adjustment (choke/expand)
    • Morphological operations for cleanup
    • Support for HSV, LAB, and RGB color spaces

    2. šŸŽ­ Studio42 Layer Composer

    Professional layer composition with full creative control

    Positioning & Transforms:

    • Direct pixel coordinates (matches other Studio42 nodes)
    • Preset positioning modes (center, corners, edges)
    • Independent X/Y scaling with aspect ratio control
    • 360-degree rotation with custom anchor points
    • Horizontal/vertical flipping

    Blend Modes:

    • Normal, Multiply, Screen, Overlay
    • Soft Light, Hard Light, Color Dodge, Color Burn
    • Darken, Lighten, Difference, Exclusion
    • Add, Subtract, Divide, Linear Burn, Linear Dodge

    Quality Controls:

    • Edge feathering for smooth transitions
    • Anti-aliasing for high-quality scaling
    • Clipping control for boundary handling
    • Fit modes (fit width/height/both, fill, stretch)

    3. šŸŽÆ Studio42 PatchLift Loader

    Interactive image loading with visual patch selection

    Visual Selection:

    • Grid overlay with adjustable size and opacity
    • Real-time coordinate display
    • Customizable selection colors and thickness
    • Shape-based selection (Rectangle, Circle, Rounded Rectangle)

    Smart Extraction:

    • True shape-based cropping with masks
    • Chroma green background for non-rectangular shapes
    • Aspect ratio maintenance options
    • Preview scaling for detailed work

    4. šŸŽÆ Studio42 PatchDrop

    Intelligent patch placement with advanced integration

    Placement Control:

    • Automatic coordinate-based placement
    • Manual position override
    • Center and preset positioning modes
    • Boundary clipping and overflow handling

    Transform & Effects:

    • Scale and rotation with quality resampling
    • Edge feathering for seamless integration
    • Multiple blend modes for creative effects
    • Opacity control with professional alpha compositing

    Mask Support:

    • Optional patch mask input for transparency control
    • Shape-aware placement masks
    • Combined mask processing for complex compositions

    5. šŸŽ„ Studio42 Video PatchLift Loader

    Advanced video loading with frame-by-frame patch selection

    Video Loading:

    • Support for MP4, WebM, MKV, AVI, MOV, GIF formats
    • Frame sampling controls (skip, select every nth)
    • Custom resolution and frame rate override
    • Batch frame processing with memory optimization

    Patch Selection:

    • Consistent patch extraction across video sequences
    • Visual overlay with frame indicators
    • Temporal coordinate display
    • NEW: Mask output for extracted patches

    6. šŸŽ„ Studio42 Video PatchDrop

    Professional video patch placement with temporal consistency

    Video Processing:

    • Frame-by-frame patch placement
    • Temporal consistency controls for smooth motion
    • Motion blur for natural integration
    • Edge blending for seamless transitions

    Advanced Features:

    • Position smoothing across frames
    • Video-optimized blend modes
    • Temporal mask consistency
    • Professional preview with visual indicators

    šŸš€ Installation

    Method 1: ComfyUI Manager (Recommended)

    1. Open ComfyUI Manager in your ComfyUI interface
    2. Search for "Studio42"
    3. Click Install
    4. Restart ComfyUI

    Method 2: Manual Installation

    # Navigate to ComfyUI custom nodes directory
cd ComfyUI/custom_nodes/

# Clone the repository
git clone https://github.com/your-repo/Studio42-ComfyUI.git

# Install dependencies
pip install -r Studio42-ComfyUI/requirements.txt

# Restart ComfyUI

    Method 3: Git Installation in ComfyUI

    # In ComfyUI custom_nodes directory
git clone https://github.com/your-repo/Studio42-ComfyUI.git

    šŸ“‹ Dependencies

    Core Requirements (Always Needed)

    Pillow>=9.5.0
opencv-python>=4.8.0
numpy>=1.24.0

    AI Background Removal (Optional - Graceful Fallback)

    # For Rembg 2.0 with GPU acceleration
pip install rembg[gpu]

# For BiRefNet models
pip install transformers>=4.30.0 timm

# For GPU support (choose one)
pip install onnxruntime-gpu>=1.15.0      # NVIDIA
pip install onnxruntime-rocm>=1.15.0     # AMD

    Video Processing (Required for Video Features)

    pip install ffmpeg-python>=0.2.0

    Performance Recommendations

    • CPU-only: All traditional methods work perfectly
    • 4GB+ VRAM: Recommended for AI background removal
    • 8GB+ VRAM: Optimal for BiRefNet HR models and batch processing
    • System FFmpeg: Install system FFmpeg for optimal video performance

    šŸŽØ Professional Workflows

    1. Advanced Background Replacement

    Image ā†’ Studio42 Background Remover (BiRefNet-General)
     ā†’ Studio42 Layer Composer (Soft Light blend)
     ā†’ Professional result with natural integration

    Settings for Best Results:

    • Method: birefnet-general or birefnet-portrait for people
    • Enable multi_pass_keying for complex edges
    • Use spill_suppression: 0.8 for color bleeding removal
    • Apply edge_softness: 2.0 for smooth transitions

    2. Selective Image Enhancement

    Image ā†’ Studio42 PatchLift Loader (select area)
     ā†’ Apply effects/filters to patch
     ā†’ Studio42 PatchDrop (overlay blend, feathered edges)
     ā†’ Seamlessly integrated enhancement

    Pro Tips:

    • Use feather_edges: 5-15 for natural blending
    • Try overlay or soft_light blend modes for enhancements
    • Enable resize_patch: false to maintain original quality

    3. Professional Chroma Key (Green Screen)

    Green Screen Image ā†’ Studio42 Background Remover

    Professional Settings:

    • Method: chroma_key_professional
    • Color: green
    • Primary threshold: 0.15 (adjust for your green screen)
    • Enable spill_suppression: 0.8
    • Enable spill_preserve_luminance: true
    • Set edge_softness: 2.0 for natural edges

    4. Video Patch Processing

    Video ā†’ Studio42 Video PatchLift Loader
     ā†’ Apply temporal effects
     ā†’ Studio42 Video PatchDrop (with temporal consistency)
     ā†’ Professional video result

    Video Settings:

    • Enable temporal_consistency: 0.3 for smooth motion
    • Use motion_blur: 1.0 for natural movement
    • Enable edge_blending: true for seamless integration

    5. Batch Background Removal

    Multiple Images ā†’ Studio42 Background Remover

    Batch Optimization:

    • Start with batch_size: 4 and adjust based on VRAM
    • Use use_fp16: true for memory efficiency
    • Consider processing_resolution: 1024 for speed vs quality balance

    āš™ļø Advanced Configuration

    Background Remover Pro Settings

    For Portraits:

    Method: birefnet-portrait
Multi-pass keying: true
Spill suppression: 0.8
Edge softness: 2.0
Color space: hsv

    For Product Photography:

    Method: chroma_key_professional
Chroma color: green (or blue)
Primary threshold: 0.12
Spill suppression: 0.9
Edge detection: true

    For Complex Scenes:

    Method: hybrid_ai_traditional
Use both AI and traditional methods
Enable garbage matte
Morphological operations: true

    Layer Composer Pro Settings

    For Realistic Compositing:

    Blend mode: soft_light or overlay
Opacity: 0.85-0.95
Feather edges: 3-8 pixels
Anti-aliasing: true

    For Creative Effects:

    Blend mode: screen, multiply, or difference
Experiment with anchor points for rotation
Use fit modes for automatic sizing

    šŸ› ļø Troubleshooting

    Common Issues

    Memory Issues:

    • Reduce batch_size parameter (try 2 or 1)
    • Enable use_fp16: true for AI methods
    • Use traditional methods for low-VRAM systems
    • Close other applications using GPU memory

    AI Models Not Loading:

    • Check internet connection (models download automatically)
    • Verify sufficient disk space (models can be 100MB+)
    • Traditional methods will activate automatically as fallback
    • Try restarting ComfyUI after first model download

    Video Processing Issues:

    • Install system FFmpeg for best compatibility
    • Check video format compatibility (MP4/WebM recommended)
    • For large videos, consider reducing resolution or frame rate
    • Use frame_load_cap to limit memory usage

    Installation Problems:

    • Ensure Python environment matches ComfyUI
    • Install dependencies in the correct virtual environment
    • Check for conflicting package versions
    • Try manual dependency installation

    Performance Optimization

    For Speed:

    • Use traditional methods (chroma/luma key)
    • Reduce processing_resolution
    • Disable advanced features for batch processing
    • Use smaller batch_size to prevent memory swapping

    For Quality:

    • Use BiRefNet models for AI processing
    • Enable multi-pass keying
    • Use higher processing_resolution
    • Enable edge detection and morphological operations

    For Memory Efficiency:

    • Enable use_fp16: true
    • Use batch processing with appropriate batch_size
    • Traditional methods use minimal memory
    • Close other GPU applications

    šŸ”§ Technical Details

    Image Format Support

    • Input: RGB, RGBA, Grayscale (auto-converted to RGB)
    • Output: RGB for images, separate alpha masks
    • Internal: Professional RGBA compositing
    • Bit Depth: 8-bit input/output, 32-bit float internal processing

    Video Format Support

    • Containers: MP4, WebM, MKV, AVI, MOV, GIF
    • Codecs: H.264, H.265, VP8, VP9, AV1 (depends on system FFmpeg)
    • Features: Frame sampling, resolution override, frame rate control

    Color Space Support

    • RGB: Standard processing mode
    • HSV: Recommended for chroma keying
    • LAB: Professional color-accurate processing
    • Auto-conversion: Seamless between color spaces

    Performance Characteristics

    • Memory Usage: Optimized for batch processing
    • GPU Acceleration: CUDA and ROCm support
    • CPU Fallback: All features work without GPU
    • Threading: Multi-threaded where applicable

    šŸ“š Examples & Tutorials

    Example Workflows

    Check the examples/ directory for sample workflows demonstrating:

    • Professional green screen removal
    • Advanced layer composition techniques
    • Video patch processing workflows
    • Batch processing setups

    Community Resources

    šŸ“„ License

    This project is licensed under the MIT License - see the LICENSE file for details.

    šŸ¤ Contributing

    Contributions are welcome! Please feel free to submit Pull Requests. For major changes, please open an issue first to discuss what you would like to change.

    Development Setup

    git clone https://github.com/your-repo/Studio42-ComfyUI.git
cd Studio42-ComfyUI
pip install -r requirements.txt -r requirements-dev.txt

    šŸ“ž Support

    For support and updates:

    šŸ™ Acknowledgments

    • ComfyUI Team for the excellent node-based interface
    • Rembg Contributors for the background removal models
    • BiRefNet Team for state-of-the-art segmentation
    • OpenCV Community for computer vision tools
    • PIL/Pillow Team for image processing capabilities

    Studio42 Image, Audio, and Video Editing Suite - Professional tools for the modern creative workflow.

    šŸŽ¬ Transform your ComfyUI workflows with professional-grade editing capabilities