Smart Image Slicer

Automatically detect and separate multiple objects from images using computer vision techniques, with intelligent background removal and transparent PNG output.

When to Use This Skill

Use this skill when users request:

• Separating sprite sheets into individual sprites
• Extracting character animation frames from sequence images
• Isolating UI elements from design mockups
• Converting hand-drawn sketches into individual digital assets
• Batch processing images with multiple objects
• Any task involving "cutting out", "separating", or "extracting" objects from images

Key Capabilities

Intelligent Object Detection

• Uses OpenCV contour detection to identify separate objects
• Filters out noise and small artifacts automatically
• Handles complex shapes and irregular boundaries

Smart Background Removal

• Analyzes local background colors for each object
• Supports gradient and non-uniform backgrounds
• Generates transparent PNG outputs
• Preserves object details while removing backgrounds

Configurable Processing

• Adjustable area thresholds to filter noise
• Tolerance settings for background removal precision
• Customizable detection sensitivity
• Preset configurations for common use cases

How to Use This Skill

1. Execute the Smart Slice Script

The core functionality is provided by scripts/smart_slice.py. Use it directly:

python3 scripts/smart_slice.py input_image.png

2. Parameter Optimization

For best results, adjust parameters based on image characteristics:

Basic Usage:

python3 scripts/smart_slice.py input.png -o output_folder

With Custom Parameters:

python3 scripts/smart_slice.py input.png -m 800 -t 45 -b 235

Parameter Guidelines:

• -m, --min-area: Minimum object size (default: 500)
  • Increase for images with many small noise elements
  • Decrease to capture smaller objects
• -t, --tolerance: Background removal sensitivity (default: 40)
  • Increase for gradient/uneven backgrounds
  • Decrease for high-contrast, uniform backgrounds
• -b, --bg-threshold: Background detection threshold (default: 240)
  • Decrease for darker backgrounds
  • Increase for pure white backgrounds

3. Use Configuration Presets

Reference assets/config_presets.json for optimized settings:

• game_sprites: For game character sprites and icons
• character_animations: For animation frame sequences
• ui_elements: For interface components and buttons
• hand_drawn: For scanned sketches and artwork
• high_precision: For detailed, precision-required images

4. Troubleshooting Common Issues

Consult references/usage_guide.md for detailed troubleshooting:

• Too many fragments: Increase min_area parameter
• Missing objects: Decrease bg_threshold and min_area
• Poor background removal: Adjust tolerance parameter
• Over-removal of object parts: Decrease tolerance parameter

Workflow Integration

For Game Development

1. Process sprite sheets to extract individual character frames
2. Use extracted sprites directly in game engines
3. Maintain transparent backgrounds for proper layering

For UI Design

1. Extract individual elements from design mockups
2. Create reusable component libraries
3. Generate assets for different screen densities

For Digital Art

1. Convert hand-drawn sketches to digital assets
2. Separate complex illustrations into layers
3. Create animation-ready sprite sequences

Technical Requirements

Ensure the following dependencies are installed:

pip install opencv-python pillow numpy

The skill works with:

• Python 3.6+
• Common image formats (PNG, JPG, BMP)
• Images of any resolution (1000x1000+ recommended)

Output Format

• Files saved as sprite_000.png, sprite_001.png, etc.
• Transparent PNG format preserving alpha channel
• Original object dimensions maintained
• Organized in specified output directory

Best Practices

Input Image Preparation

• Use solid, preferably white backgrounds for best results
• Ensure good contrast between objects and background
• Higher resolution images generally produce better results

Parameter Tuning Strategy

1. Start with default parameters
2. Analyze initial results
3. Adjust parameters based on specific image characteristics
4. Use preset configurations for similar image types

Quality Assurance

• Review extracted objects for completeness
• Check transparency edges for artifacts
• Verify all intended objects were detected
• Manually adjust parameters if needed for optimal results