Harmony-Rule Engine

An interoperability layer between Lampang natural-dye color knowledge and contemporary design practice — CIELAB-based
v2.0 · Dataset: Netithammakorn, Sakon & Shen (2025), Color Research & Application

1. Classify a Palette

Select 2 or more colors from the 33-color natural dye dataset. The engine will classify which harmony category (or categories) best match your selection.
All (33) Burma Padauk Siamese Senna Indian Almond Eucalyptus Lac

Your Selection

Click × on a color chip to remove it.

2. Generate a Palette

Pick a harmony category. The engine will sample colors from the dataset that satisfy the rule.

3. Export Palette

Current palette can be exported as HEX list, CMYK table, or CSS variables.
Design Tokens (W3C DTCG) format is directly consumable by Figma token plugins, Style Dictionary, and compatible design-system pipelines. 

    

  


  
  

    

      
About this prototype

      

        This web application is the interoperability layer accompanying the research paper
        "A Harmony-Rule Engine as an Interoperability Bridge for Natural Dye Color Systems"
        (target: Color Research & Application, Wiley). It translates craft-derived color knowledge into interoperable design assets — classify, generate, and export categorised palettes drawn from a measured natural-dye dataset.
      


      
The Six Harmony Rules

      

        Rule definitions derived from the curated reference set. Thresholds are designer-editable and were calibrated against expert-curated pairs.
      

      


      
Dataset provenance

      

        The 33-color dataset is derived from:
        

        Netithammakorn, N., Sakon, T. & Shen, Y. (2025). A study of natural dyes to create color palette for creative design. Color Research & Application. DOI: 10.1002/col.70022
        


        Colors were measured using a Datacolor Spectro 700V spectrophotometer under a D65 illuminant and the CIE 1964 10° standard observer, with four replicate measurements averaged to produce final CIELAB (L*, a*, b*) coordinates.
      


      
Technology

      

        Vanilla HTML/CSS/JavaScript. All color math is implemented directly — no external library dependencies — so the application runs in any modern browser without a server.
      


      
Known limitations

      
    
        
  • Rule thresholds are calibrated from the curated reference pairs; further tuning may be appropriate when extending the framework to other regional dye palettes.
    • 
        
  • Confidence scoring is a rule-match closeness metric, not a calibrated probability.
    • 
        
  • Measured CIELAB values are rendered as approximate sRGB for display; the underlying L*a*b* coordinates remain the authoritative reference.
    •