Deep Learning-Based Multi-Class Classification of Cough Sounds for Respiratory Disease Detection and Real-Time Screening
Bibliographic record
Abstract
Respiratory illnesses like COVID-19, tuberculosis, and other lung conditions often show distinct cough patterns. Due of this, analyzing cough sounds is useful. It provides a non-invasive way to diagnose these conditions. This study introduces a deep learning framework that can classify cough sounds into seven categories which are: healthy, no_resp_illness_exposed, positive_asymp, positive_mild, positive_moderate, recovered _ full, and resp_illness_notjdentified. The dataset includes 2,277 confirmed real cough recordings. We processed the data by combining metadata, validating audio, and extracting features using Mel-Frequency Cepstral Coefficients (MFCCs). We trained a hybrid CNN-LSTM model, in which an overall accuracy of 55% and a weighted Fl-score of 0.50 is achieved. The healthy category had the best performance, reaching an F1 score of 0.73. While the classification accuracy for minority classes was moderate, the results show that cough acoustics offer significant diagnostic potential. To enhance user experience, we launched the trained model in a web application based on Streamlit. This lets users upload a cough recording or record audio directly. Predictions and confidence scores are available within seconds. This setup connects algorithm research to practical healthcare solutions. It provides a cost-effective, privacy-respecting, and an accessible tool for initial screening of respiratory illnesses. Future improvements will focus on balancing the dataset, enhancing key features, and implementing federated learning for privacy-conscious clinical use along with improved accuracy and confidence levels.
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 | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
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".