Releases · shawntz/charisma

Initial CRAN Release

First public release of charisma - an R package for reproducible color characterization of digital images for biological studies (and other things).

Automatic Color Classification: Classify image colors into one of 10 discrete, biologically-inspired categories (black, blue, brown, green, grey, orange, purple, red, white, yellow)
CLUT System: Color Look-Up Table (CLUT) with scientifically-defined HSV boundaries for each color category, validated across 3.6+ million color coordinates
Main Functions:
- charisma(): Main color classification pipeline with automatic thresholding
- charisma2(): Rewind and edit saved charisma objects without re-running full pipeline
- color2label(): Convert individual RGB/HSV colors to discrete labels
- validate(): Validate CLUT completeness across color space
Integration:
- Leverages recolorize for image preprocessing and color segmentation
- Optional pavo integration for color pattern geometry analysis
- Supports both interactive and automated workflows
Reproducibility:
- Complete state tracking with merge/replacement history
- Save and reload analyses with full provenance
- Deterministic color classification for cross-study comparisons
Visualization:
- Plot methods for charisma objects
- Mosaic plots showing color proportions
- Color masks and classification overlays