Beyond Lux thresholds: a systematic pipeline for classifying biologically relevant light contexts from wearable data

TL;DR

This paper presents a reproducible spectral-temporal data pipeline for classifying natural versus artificial light from wearable spectrometer data, demonstrating high accuracy and establishing design rules for generalizable contextual light classification.

Contribution

The authors develop and validate a subject-wise, reproducible pipeline with specific processing steps and design rules for classifying light contexts from wearable spectral data, advancing beyond simple illuminance thresholds.

Findings

Achieved AUC = 0.938 for natural vs. artificial classification

Indoor vs. outdoor classification remains challenging with AUC ~0.75

Spectral-temporal modeling outperforms threshold-based methods

Abstract

Background: Wearable spectrometers enable field quantification of biologically relevant light, yet reproducible pipelines for contextual classification remain under-specified. Objective: To establish and validate a subject-wise evaluated, reproducible pipeline and actionable design rules for classifying natural vs. artificial light from wearable spectral data. Methods: We analysed ActLumus recordings from 26 participants, each monitored for at least 7 days at 10-second sampling, paired with daily exposure diaries. The pipeline fixes the sequence: domain selection, log-base-10 transform, L2 normalisation excluding total intensity (to avoid brightness shortcuts), hour-level medoid aggregation, sine/cosine hour encoding, and MLP classifier, evaluated under participant-wise cross-validation. Results: The proposed sequence consistently achieved high performance on the primary task, with representative configurations reaching AUC = 0.938 (accuracy 88%) for natural vs. artificial classification on the held-out subject split. In contrast, indoor vs. outdoor classification remained at feasibility level due to spectral overlap and class imbalance (best AUC approximately 0.75; majority-class collapse without contextual sensors). Threshold baselines were insufficient on our data, supporting the need for spectral-temporal modelling beyond illuminance cut-offs. Conclusions: We provide a reproducible, auditable baseline pipeline and design rules for contextual light classification under subject-wise generalisation. All code, configuration files, and derived artefacts will be openly archived (GitHub + Zenodo DOI) to support reuse and benchmarking.