Mapping to the PICO-8 palette, perceptually
4 days ago
- #image processing
- #color spaces
- #computer graphics
- Explores mapping image pixels to a palette using different color spaces.
- Compares Euclidean distance in sRGB, CAM16-UCS, Oklab, and CIELAB with luminance weighting.
- CAM16-UCS results were not better than Oklab or weighted CIELAB, but all outperformed sRGB.
- Discusses the impact of viewing conditions on CAM16-UCS results.
- Highlights the limitations of ignoring image spatial structure in color mapping.
- Includes Python code examples for color space conversions and distance calculations.
- Mentions the Helmholtz-Kohlrausch effect affecting luminance-based mappings.
- Provides a reference to a book on computer graphics and experiment result images.