# Spatiotemporal Feature Learning for Daily-Life Cough Detection Using FMCW Radar

## TL;DR

This paper introduces a radar-based system that accurately detects coughs in daily life using a deep learning model combining spatial and temporal features.

## Contribution

The novel framework integrates CNN and Self-Attention mechanisms for robust cough detection using FMCW radar data.

## Key findings

- The model achieved a mean F1-score of 0.974±0.016 under subject-independent cross-validation.
- High accuracy (99.05±0.55%) and specificity (99.73±0.23%) demonstrate robust performance in real-world scenarios.

## Abstract

Cough is a key symptom reflecting respiratory health, with its frequency and pattern providing valuable insights into disease progression and clinical management. Objective and reliable cough detection systems are therefore of broad significance for healthcare and remote monitoring. However, existing algorithms often struggle to jointly model spatial and temporal information, limiting their robustness in real-world applications. To address this issue, we propose a cough recognition framework based on frequency-modulated continuous-wave (FMCW) radar, integrating a deep convolutional neural network (CNN) with a Self-Attention mechanism. The CNN extracts spatial features from range-Doppler maps, while Self-Attention captures temporal dependencies, and effective data augmentation strategies enhance generalization by simulating position variations and masking local dependencies. To rigorously evaluate practicality, we collected a large-scale radar dataset covering diverse positions, orientations, and activities. Experimental results demonstrate that, under subject-independent five-fold cross-validation, the proposed model achieved a mean F1-score of
0.974±0.016
and an accuracy of
99.05±0.55
%, further supported by high precision of
98.77±1.05
%, recall of
96.07±2.16
%, and specificity of
99.73±0.23
%. These results confirm that our method is not only robust in realistic scenarios but also provides a practical pathway toward continuous, non-invasive, and privacy-preserving respiratory health monitoring in both clinical and telehealth applications.