FFT Spectrum Spread With Machine Learning (ML) Analysis of Triaxial Acceleration From Shirt Pocket and Torso for Sensing Coughs While Walking
Bibliographic record
Abstract
Early detection of respiratory distress, marked by coughing associated with pandemics such as Covid, severe acute respiratory syndrome, and influenza, has become important for early public health preparedness. Recognizing respiratory distress from data pooled from accelerometers and other sensors common in phones/wearables can be a useful tool in tracking diseases in larger populations. However, detecting low-/medium-intensity coughs, which are a precursor to influenza/Covid, are harder to detect in the presence of human activity especially walking. In this letter, we study spectrum-spread features of triaxial accelerometer signals measured from the human torso during coughs. In particular, we analyze the vestigial sideband like spurs that cough-induced motion of the torso produces alongside walking signal between 0.2 and 2 Hz and propose the use of its spectral spread square metric in discerning coughs during walking action in test subjects of different sizes. Unlike prior works on time-domain measurements or spectral summation (units: g) in multiple bands, this work uses bandwidth, i.e., spectrum-spread features of acceleration signals (units: Hz2) to detect low to medium intensity coughs from a single accelerometer worn on the chest or shirt pocket or stomach. Acceleration signals measured at these points in five test subjects of varying heights, age, and weight show its median square spectral spread increase prominently alongY(up-down) andZaxes (front-back) from between 0.016–0.0167 Hz2to between 0.023–0.026 Hz2with a cough-detection threshold observed at 0.02 Hz2for all axes. Using a machine learning (ML) classification model with these spectral spread features results in cough detection accuracy of 92.5, 92.2, and 91.5% with k-nearest neighbors (kNN), and 94.3, 96.1, and 93.6% using Support Vector Machine (SVM) ML models for all three torso points especially shirt pocket where phones are commonly worn.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".