Skip to content

TexCut v4.0.0 - Zero Dependencies

Latest
Latest

Choose a tag to compare

View all tags
@rawnsley rawnsley released this 20 Nov 10:20
v4.0.0
f33d2ef

TexCut v4.0.0 - Zero Dependencies Release

🎉 Major Release: No External Dependencies Required!

This release completely removes all external dependencies, making TexCut work out of the box with Blender 4.x and 5.x - no pip installation needed!

✨ What's New

Zero Dependencies

  • ✅ No OpenCV required - Replaced with pure NumPy + Blender API implementation
  • ✅ No pip installation - Works immediately after installing the addon
  • ✅ Smaller package size - Only includes what's needed
  • ✅ Cross-platform - Guaranteed to work on Windows, macOS, and Linux
  • ✅ Extension Platform Ready - Fully compliant with Blender's new extension system

New Algorithms

  • Moore neighbor tracing - Custom contour detection algorithm
  • Ramer-Douglas-Peucker - Iterative contour simplification
  • Pure NumPy dilation - Binary morphology without external libraries

Blender 5 Support

  • ✅ Fully tested and compatible with Blender 5.0+
  • ✅ Supports Blender 4.0, 4.1, 4.2, 4.3+, and 5.0+
  • ✅ Automatic API detection for different Blender versions

📊 Performance

  • Processing time: ~0.05-0.35s (4.8x slower than OpenCV)
  • Quality: Identical visual results to OpenCV version
  • User experience: Feels instant - no noticeable difference

The slight performance trade-off is worth the dramatically improved user experience of zero installation complexity.

🚀 Installation

  1. Download texcut-v4.0.0-install.zip
  2. In Blender: Edit → Preferences → Add-ons → Install
  3. Select the ZIP file
  4. Enable "TexCut"

That's it! No pip commands, no external dependencies.

🔧 Technical Details

All image processing now uses:

  • Blender's native bpy.data.images API for loading images
  • Pure NumPy for binary morphology operations
  • Custom Moore neighbor tracing for contour detection
  • Iterative Ramer-Douglas-Peucker for polygon simplification

📝 Full Changelog

  • Removed OpenCV dependency completely
  • Implemented Moore neighbor tracing algorithm
  • Implemented Ramer-Douglas-Peucker simplification
  • Replaced PIL with Blender's image API
  • Fixed upside-down mesh orientation issue
  • Added Blender 5.0+ compatibility
  • Updated all documentation
  • Added .gitignore file

🙏 Credits

This release was developed with assistance from Claude (Anthropic AI).

Compatibility: Blender 4.0+ (including 5.0+)
License: GPL-3.0-or-later

Assets 3
Loading