Dithering and Stippling

This directory contains a collection of Python tools for generating stippled images using Lloyd's relaxation and multi-tone color stippling algorithms.

Core Modules

1. Lloyd.py

Implements algorithms for point distribution and relaxation:

• Ordinary Lloyd Relaxation: Standard Centroidal Voronoi Tessellation (CVT) update for uniform point distribution.
• Modified Lloyd Relaxation: CVT update weighted by a density function $\rho(x, y)$, allowing points to cluster in regions of higher density (e.g., darker areas of an image).

2. Stippling.py

Defines the BaseStipple class for image processing:

• Image Input: Loads and normalizes images to $[0, 1]$ range.
• RGB Interpolators: Uses scipy.interpolate.RegularGridInterpolator for efficient retrieval of color data at arbitrary coordinates.
• HSV & Density Mapping: Methods to create interpolators based on hue, saturation, value, and calculated color density.

3. Stippling2.py

Extends the base functionality with the MilisecondsStipple class, implementing algorithms from the paper "Milliseconds Color Stippling" (MM '21) by Lei Ma, Jian Shi, and Yanyun Chen:

• Forward Multi-tone Color Stippling: Optimized for placing dark stipples on a white background.
• Reverse Multi-tone Color Stippling: Optimized for placing light stipples on a black background, preventing overcrowding in dark regions.

4. Stippling3.py

Introduces the SequentialStipple class, refactored for sequential and multi-toned dithering tasks utilizing customizable, pre-defined dome libraries (palettes).

• Dynamic Density Mapping: Features options for generating density maps from luminance, inverse luminance, grayscale, or uniformly.
• Color Assignment & Error Diffusion: Includes multiple quantization methods (arithematic, area_weighted, and Floyd-Steinberg) to intelligently allocate palette colors onto unstructured scattered Voronoi cells while diffusing color discrepancies contextually.

Usage and Examples

• demo.ipynb: Interactive demonstration of Lloyd relaxation with both synthetic (Gaussian) and image-based density functions.
• example_stippling2.py: A complete script demonstrating the application of forward and reverse stippling on a color image, including visualization.
• example_stippling3.py: A comprehensive script executing the sequential multi-toned stippling workflow. Demonstrates density-based rejection sampling alongside Voronoi-based color assignment via advanced dithering methods like Floyd-Steinberg error diffusion.

Requirements

• numpy
• scipy
• Pillow
• matplotlib