Skin-Tone-Invariant Topological Signal Processing: A Framework for Bias-Reducing Optical Measurement Systems

Photoplethysmography (PPG) is an optical method for measuring changes in blood volume in the heart and forms the foundation of wearable cardiac monitoring devices. However, this method systematically fails in people with darker skin tones due to melanin absorption. Researchers presented the Melanin Absorption Invariance (MAI) framework, which significantly reduces bias across different skin tones in cardiac feature extraction. The method preserves the topological structure of the signal rather than its geometric properties. The research was validated on 656 recordings from 33 subjects with Fitzpatrick skin types III to VI using a Samsung Galaxy mobile phone under various lighting conditions and movement types. Results demonstrate substantial bias reduction across all skin tone groups, with the largest effects observed in Fitzpatrick IV and VI populations, which are most affected by current systems.