From: The coming era of a new auscultation system for analyzing respiratory sounds
Topic of study | Number of subject/recording | Number of classes | Applied model | Result | References | |
---|---|---|---|---|---|---|
Recognition of pulmonary diseases from lung sounds using CNN and LSTM | 213/1483 | 6: Normal, asthma, pneumonia, bronchiectasis, COPD, heart failure | CNN, biLSTM | Accuracy | [12] | |
biLSTM: 98.16% | ||||||
CNN: 99.04% | ||||||
CNN + biLSTM: 99.62% | ||||||
Classification of respiratory sounds using OST and deep residual networks | Not available/489 | 3: Crackle, wheeze, normal | OST, ResNets | Accuracy | [13] | |
STFT: 93.98% | ||||||
ST: 97.79% | ||||||
OST + ResNets: 98.79% | ||||||
Detection of respiratory sounds based on wavelet coefficients and machine learning | 130/705 | 3: Crackles, rhonchi, normal | SVM, ANN, KNN | Accuracy | [14] | |
SVM: 69.50% | ||||||
ANN: 85.43% | ||||||
KNN: 68.51% | ||||||
Benchmarking of eight recurrent neural network variants for breath phase and adventitious sound detection | 279/9765 | 6: Inhalation, exhalation, wheeze, stridor, rhonchus, crackles | LSTM, GRU, BiLSTM, BiGRUc, CNN-LSTM, CNN-GRU, CNN-BiLSTM, and CNN-BiGRU | F1 score | [15] | |
LSTM: 73.9% | GRU: 77.6% | |||||
BiLSTM: 76.2% | BiGRU: 78.4% | |||||
CNN-LSTM: 78.1% | CNN-GRU: 80.6% | |||||
CNN-BiLSTM: 80.3% | CNN-BiGRU: 80.6% | |||||
Classification of lung sounds through DS-CNN models with fused STFT and MFCC features | Not available/12,691 | 4: Normal, wheeze, crackle, unknown | DS-CNN, VGG-16, AlexNet, DS-AlexNet, LSTM, GRU, TCNd | Accuracy | [16] | |
DS-CNN: 85.74% | VGG-16: 85.66% | |||||
AlexNet: 79.92% | DS-AlexNet: 80.86% | |||||
LSTM: 76.92% | GRU: 78.50% | |||||
TCN: 75.51% | ||||||
Implementation of AI algorithms in pulmonary auscultation examination | 50/522 | 4: Wheezes, rhonchi, fine and coarse crackles | Neural network | F1-score (%) | [17] | |
Coarse crackles: Doctors (42.8%), NN (47.1%) | ||||||
Fine crackles: Doctors (51.1%), NN (64.6%) | ||||||
Wheezes: Doctors (61.8%), NN (66.4%) | ||||||
Rhonchi: Doctors (61.0%), NN (72%) | ||||||
AI accuracy in detecting pathological breath sounds in children | 25/192 | 2: Crackles, wheeze | Neural network | PPAa/NPAb | [18] | |
Crackles: Clinicloud 0.95/0.99, Littman 0.82/0.96 | ||||||
Wheezes: Clinicloud 0.90/0.97, Littman 0.80/0.95 | ||||||
Classification of lung sounds using CNN | 1630/17,930 | â‘ 2: Healthy, pathological | CNN, SVM | Accuracy (CNN/SVM) | [19] | |
â‘¡ 3: Rale, rhonchus, normal | â‘ 86%/86% | â‘¡ 76%/75% | ||||
Feature extraction technique for pulmonary sound analysis based on EMD | 30/120 | 3: Normal, wheezes, crackles | ANN, SVM, GMM | Accuracy | [20] | |
EMD with ANN: 94.16% | ||||||
EMD with SVM: 93.75% | ||||||
EMD with GMM: 88.16% | ||||||
Application of deep Learning to classify the severity of COPD | Not available/120 | 2: Crackles, wheeze | NN, DBN | Accuracy | [21] | |
NN: 77.60% | ||||||
DBN: 95.84% | ||||||
Application of deep Learning to detect early COPD | 50/600 | 1: Wheeze | DBN | Accuracy | [22] | |
DBN: 93.67% | ||||||
Application of semi-supervised deep learning to lung sound analysis | 284/11627 | 2: Crackles, wheeze | SVM | AUC | [23] | |
Crackle: 0.74 | ||||||
Wheeze: 0.86 |