BreathSense: Synchronized Radar and Respiratory Belt Data for 108 Participants

BreathSense provides 432 paired recordings of millimeter-wave radar and respiratory belt data. The dataset includes data from 108 participants across four activity conditions: rest, walk, run, and stair climbing. It was created by BreathSense and last updated on Hugging Face in May 2026.

Use Cases

• Developing non-contact respiratory rate estimation algorithms based on synchronized radar and belt data.
• Training machine learning models for activity-specific respiratory pattern recognition.
• Benchmarking signal processing techniques for extracting physiological signals from mmWave radar data.
• Studying the impact of physical activity on respiratory waveforms.

Strengths

• Contains 432 paired recordings, providing a substantial sample size.
• Includes data from 108 unique participants, suggesting diversity in the subject pool.
• Covers four distinct physical activity conditions, enabling comparative analysis.
• Features synchronized ground-truth respiratory waveforms for validation.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and file formats are unknown, which may limit suitability assessment.
• Data may reflect participant or collection bias inherent to the source.

Provenance

Source

BreathSense

Collection Method

Collected using a 77 GHz FMCW radar and synchronized respiratory belts from participants performing scripted activities.

Freshness

Last updated 2026-05-07 09:33:52; freshness should be verified.