Thermal Imaging and Radar for Remote Sleep Monitoring of Breathing and Apnea

This repository contains the code used in the paper "Thermal Imaging and Radar for Remote Sleep Monitoring of Breathing and Apnea", authored by Kai Del Regno, Alexander Vilesov, Adnan Armouti, Anirudh Bindiganavale Harish, Selim Emir Can, Ashley Kita, and Achuta Kadambi.

Abstract

Polysomnography (PSG) is the gold standard for diagnosing sleep disorders, but it is cumbersome and expensive. At-home sleep apnea testing (HSAT) solutions are contact-based, which may not be tolerated by some patient populations. This paper presents the first comparison of radar and thermal imaging for non-contact sleep monitoring. Our results show that thermal imaging outperforms radar in detecting apneas. Additionally, we propose a multimodal method to classify obstructive and central sleep apneas using synchronized radar and thermal data.

We also provide the code, hardware setup, and circuit schematics for synchronized radar and thermal data collection.

Contributions

1. A comparison of radar and thermal modalities for remote sleep monitoring and detection of breathing and sleep apnea.
2. A non-contact multimodal thermal and radar stack for the detection and clinically relevant classification of sleep apnea.
3. A dataset composed of 10 sleeping patients with hardware-synchronized thermal videos, frequency-modulated continuous wave (FMCW) radar data, and ground truth waveforms and annotations by a certified sleep technician at a sleep center. We also open-source our data-collection framework, code base, and circuit schematics for collecting hardware-synchronized radar and thermal data.

Repository Contents

1. data_collection/: Contains all code, circuit diagrams, and instructions for capturing data from the thermal and radar sensors.
2. preprocessing/: After data collection, the data requires preprocessing before moving to the inference stage. The associated instructions for using our preprocessing library are located in the README within "preprocessing/mmProc".
3. inference/Apnea_Detection_Code/: Contains the scripts apnea_detection.py and breathing_detection.py, along with accompanying code for running breathing and apnea estimation.

Citation

If you use this code or dataset in your work, please cite:

@article{delregno2024thermal,
  title={Thermal Imaging and Radar for Remote Sleep Monitoring of Breathing and Apnea},
  author={Del Regno, Kai and Vilesov, Alexander and Armouti, Adnan and Harish, Anirudh Bindiganavale and Can, Selim Emir and Kita, Ashley and Kadambi, Achuta},
  journal={arXiv preprint arXiv:2407.11936v2},
  year={2024}
}