Context-Adaptive Color Optimization for Web Accessibility: Balancing Perceptual Fidelity and Functional Requirements
By: Lalitha A R
Potential Business Impact:
Makes websites easier for everyone to read.
We extend our OKLCH-based accessibility optimization with context-adaptive constraint strategies that achieve near-universal success rates across diverse use cases. Our original strict algorithm reached 66-77% success by prioritizing minimal perceptual change ($ΔE \leq 5.0$), optimizing for enterprise contexts where brand fidelity is paramount. However, this one-size-fits-all approach fails to serve the broader ecosystem of web developers who need accessible solutions even when strict perceptual constraints cannot be satisfied. We introduce recursive optimization (Mode~1) that compounds small adjustments across iterations, achieving 93.68% success on all color pairs and 100% success on reasonable pairs (contrast ratio $ρ> 2.0$), representing a +27.23 percentage point improvement. A relaxed fallback mode (Mode~2) handles pathological edge cases, reaching 98.73% overall success. Evaluation on 10,000 realistic web color pairs demonstrates that context-aware constraint relaxation, combined with absolute hue preservation, enables practical accessibility compliance while maintaining brand color identity. The median perceptual change remains zero across all modes (most pairs already comply), while the 90th percentile reaches $ΔE_{2000} = 15.55$ in Mode~1 -- perceptually acceptable when hue invariance preserves the essential character of the original color. The approach is deployed in CM-Colors v0.5.0 (800+ monthly downloads), providing developers with explicit control over the accessibility-fidelity trade-off appropriate to their context.
Similar Papers
Perceptually-Minimal Color Optimization for Web Accessibility: A Multi-Phase Constrained Approach
Human-Computer Interaction
Makes websites easy to read without changing colors.
Chameleon: Automated Color Palette Adaptation for Dark Mode Data Visualizations
Human-Computer Interaction
Makes charts look good in dark mode.
A Perceptually Inspired Variational Framework for Color Enhancement
CV and Pattern Recognition
Improves picture colors to look more natural.