Pith. sign in

REVIEW 3 major objections 5 minor 31 references

Disentangling Dual-Encoder Masked Autoencoder for Respiratory Sound Classification

T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A dual-encoder masked autoencoder that separates disease-related from disease-irrelevant sound features improves respiratory sound classification and reduces the need for domain labels.

desk verdict A plausible small win on ICBHI from combining Audio-MAE with Siamese and vCLUB losses, but the paper's central domain-mismatch claim is not tested by its random-split experiments. read the letter →

arxiv 2506.10698 v2 pith:TPAHFTZM submitted 2025-06-12 eess.AS cs.SD

Peidong Wei , Shiyu Miao , Lin Li This is my paper · ORCID
classification eess.AScs.SD
keywords respiratorysoundclassificationfeaturedisentanglementmaskedautoencoderself-supervisedlearningdomainadaptationICBHIdatasetmutualinformationminimizationSiameseloss
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper proposes a modified masked autoencoder, DDE-MAE, for classifying respiratory sounds, and argues that forcing the model to split its representation into two independent streams improves accuracy when recordings come from many stethoscopes, patients, and environments. The disease-related encoder works on the original spectrogram, while the disease-irrelevant encoder is trained to be invariant to random shuffling of the time axis, so it must rely on static, background-like content. A mutual-information penalty keeps the two streams independent. On the ICBHI respiratory sound benchmark, the full model reaches an average score of 61.50%, beating reconstruction-only fine-tuning by 1.57 points and reaching the highest sensitivity (53.69%) among the compared methods. This matters because the method promises domain adaptation without any labels for device, patient, or environment.

What carries the argument

The central mechanism is a dual-encoder masked autoencoder in which one encoder ($E_r$) learns disease-related content from the intact spectrogram and the other ($E_i$) learns disease-irrelevant content by seeing both the original and a time-shuffled version of the spectrogram. Time shuffling randomly reorders the time axis, destroying temporal dependencies while keeping spectral content, and a Siamese loss forces $E_i$ to produce similar embeddings for the original and shuffled inputs. Independence between the two streams is enforced by minimizing the vCLUB upper bound on their mutual information. Both encoders feed one decoder that reconstructs masked spectrogram patches, so the representation pair must jointly retain enough information to rebuild the input. After this disentanglement stage, only the disease-related encoder is fine-tuned for the classification task.

What would settle it

A direct test is to train only the disease-irrelevant encoder on the same ICBHI classes and measure its classification accuracy after disentanglement: if it stays far above chance, then time shuffling did not strip out disease information. A quantitative version is to estimate the mutual information between the time-shuffled spectrogram and the disease label; if that estimate is not near zero, the assumed separation fails.

Watch

Extended reading notes

Core claim

The paper's discovery claim is that feature disentanglement—rather than extra reconstruction training or domain supervision—is what drives the improvement. In their ablation, the full dual-encoder setup with both Siamese and mutual-information losses reaches an average score of 61.50%, against 59.93% for the same model fine-tuned only to reconstruct before classifying. The ablation also shows both losses contribute: removing the mutual-information loss drops the average score to 61.30%, and removing the Siamese loss drops it to 60.14%. The authors additionally report that the disease-irrelevant encoder, when probed with a linear classifier, predicts device and patient labels far better than the disease-related encoder does, while the disease-related encoder remains relatively focused on disease cues. This is presented as evidence that the two streams have indeed separated confounding factors from diagnostic content.

Load-bearing premise

The load-bearing premise is that randomly shuffling the order of time frames in the spectrogram wipes out the information a doctor would use to hear disease—temporal patterns like crackles and wheezes—while leaving the static spectral content of devices and patients intact, so a network forced to ignore the shuffle learns only non-disease features.

Editorial extensions

If this is right

  • If the claim holds, respiratory sound classifiers can be made more robust to new stethoscopes and recording environments without collecting any domain labels, because the confounding factors are absorbed by the disease-irrelevant encoder.
  • The disentanglement stage adds a real gain beyond reconstruction: the full setup beats reconstruction-only fine-tuning by 1.57 points in average score, so the separation mechanism, not just extra data exposure, is responsible for the improvement.
  • The method reaches the highest reported sensitivity among compared approaches (53.69%), which matters for screening because missing abnormal breath sounds is costlier than over-flagging.
  • The disease-related encoder's residual sensitivity to device and patient factors, acknowledged by the authors, suggests that further refinement of the separation could yield still larger gains.
  • Since the approach needs no domain labels and no extra modalities, it can be applied directly to existing respiratory sound collections, including small ones, starting from a pretrained Audio-MAE.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The same time-shuffling invariance could be transferred to other temporal audio diagnostics, such as heart sounds, bowel sounds, or sleep sounds, where the distinction between transient events and steady background also separates disease from recording conditions, though that transfer is untested here.
  • The premise that temporal shuffling preserves non-disease content is likely to be only partially true: crackles are transient, so some disease information is temporal, and a natural alternative design might pair time shuffling with frequency masking to define the irrelevant stream.
  • The device and patient probe result implies a practical diagnostic use the paper does not develop: the disease-irrelevant encoder could serve as a domain fingerprint for flagging out-of-distribution recordings.
  • Because the gap between direct training and AudioSet pretraining is large, the marginal value of disentanglement may depend on the pretrained initialization, and could shrink or grow under other pretraining recipes.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper proposes DDE-MAE, a dual-encoder masked autoencoder for respiratory sound classification. One encoder (E_R) is trained on original spectrograms to capture disease-related features; a second encoder (E_I) is trained with a Siamese loss to be invariant to random time-shuffling of the spectrogram, with the intent of capturing disease-irrelevant static content. A vCLUB-based mutual information loss encourages the two encoders' embeddings to be independent. The method is evaluated on the ICBHI dataset under the official random 60/40 split. On the primary metric AS (average of sensitivity and specificity), the full system reaches 61.50, versus 59.93 for reconstruction-only fine-tuning (S2) and 57.59 for direct classification fine-tuning (B2). An ablation shows that removing either the Siamese loss or the MI loss degrades performance. The qualitative analysis in Fig. 2 probes both encoders on device and patient labels, reporting that E_I captures these domain labels more strongly than E_R.

Significance. If the disentanglement claim were fully validated, the method would be a useful self-supervised contribution to respiratory sound classification under domain shift, since it does not require domain labels. The experimental design is generally sound: the comparison S3 versus S2 isolates the effect of the disentanglement losses from additional reconstruction training, and both losses appear to contribute. The reported improvement is modest but consistent with confidence intervals. However, the central claim in the abstract and introduction—that the method reduces domain mismatch and improves domain adaptation—is not actually tested by the experimental protocol, and the disentanglement evidence is incomplete. These are load-bearing gaps rather than presentation issues.

major comments (3)
  1. [Section 2.1 and Section 3.5] The core premise that time-shuffling the spectrogram preserves only disease-irrelevant content is not verified. Section 2.1 states that shuffling 'disrupts temporal dependencies while preserving spectral content,' but adventitious respiratory sounds have strong static spectral signatures: wheezes are near-stationary tonal components and crackles have characteristic spectral envelopes. A shuffled spectrogram can therefore still contain disease-discriminative information. Making E_I invariant to shuffling via the Siamese loss does not guarantee that E_I is disease-irrelevant. The only disentanglement evidence is Fig. 2, which linear-probes E_I and E_R on device and patient labels. It never probes E_I on the respiratory disease labels. If E_I classifies disease above chance, the claimed separation is incomplete, and the vCLUB MI loss may be discarding shared signal rather than removing domain-only content. In addition, patient labels correlate with disease state, so high patient accuracy on E_I is not by itself evidence of domain-only encoding. A disease-label linear probe on E_I (and on the shuffled-input representation) is needed to support the disentanglement claim.
  2. [Section 3.1 and Section 3.6] The paper's central claim of reducing domain mismatch and enabling generalization to unseen domains is not tested. All experiments use the official random 60/40 split of ICBHI described in Section 3.1, so train and test recordings likely share stethoscope types, patients, and recording environments. The abstract and introduction motivate the work by domain mismatch from different stethoscopes, patient demographics, and environments, and Section 3.6 argues the method does not need domain labels; however, no experiment evaluates the method on unseen stethoscopes, unseen patients, or unseen recording conditions. A cross-domain evaluation, such as leave-one-stethoscope-out or a patient-disjoint split, is required to substantiate the domain-adaptation claim. Without it, the reported improvements could reflect better in-distribution representation learning rather than domain-mismatch reduction.
  3. [Section 3.6 and Table 3] The state-of-the-art comparison is reported without confidence intervals or significance tests for the baseline methods, and the interpretation leans heavily on sensitivity. DDE-MAE has the highest SE (53.69) but the lowest or near-lowest SP (69.32) among the top-tier methods, with AS (61.50) within the range of AST + patch-mix CL (62.37) and SG-SCL (61.71). Since the primary metric is the average of SE and SP, a high-SE/low-SP operating point is not necessarily better; a sensitivity-focused conclusion needs a stated clinical or evaluation rationale and ideally a paired significance test or confidence intervals for the baselines.
minor comments (5)
  1. [Table 2] Table 2 contains two rows both labeled '- MI Loss' (69.11 and 67.61). The second row likely denotes removal of both losses or a different configuration; the label should be corrected so the ablation is unambiguous.
  2. [Section 2.2, after Eq. (3)] The text says 'the disease-related representation Z_R and the disease-relevant representation Z_I'; for consistency with the rest of the paper, Z_I should be called disease-irrelevant.
  3. [Section 3.3] The heading 'Traning details' contains a typo; it should be 'Training details'.
  4. [Section 3.3 and Section 2.3] The paper does not specify the spectrogram parameters (FFT size, hop length, window function, number of mel bands) or the MAE masking ratio and patch size used in the experiments. These details are needed for reproducibility.
  5. [Section 3.4, after Table 1] The sentence 'the comparisons between B1 and B2, as well as between S1 and S3 reveal that pretraining plays a pivotal role' is supported by the table, but the magnitude of improvement is stated for S1 over B1 and S3 over B2; the corresponding confidence intervals overlap for some metrics, so the language should be softened to avoid overstating significance.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper's quantitative claims are anchored in external ICBHI benchmarks and standard losses, with no fitted parameter renamed as a prediction.

full rationale

The derivation chain is self-contained with respect to the paper's central claims. The classification improvements in Table 1 are measured against external or controlled configurations: S3 (disentanglement fine-tuning) is compared with S2 (reconstruction-only fine-tuning), so the gain is not an artifact of merely exposing the model to ICBHI data, and Table 3 compares against published methods. The vCLUB mutual-information loss and Siamese loss are standard objectives, not quantities fitted to the reported AS/SP/SE metrics. The paper cites prior work such as Masked Autoencoders, Audio-MAE, and CLUB, but none of these citations are self-citations by the present authors, and no load-bearing premise is justified only by a citation from the same group. The time-shuffling assumption—that shuffling the time axis disrupts temporal dependencies while preserving spectral content—is an empirical premise that may be incorrect for stationary adventitious sounds like wheezes, but that is a correctness risk, not circularity: the paper does not define the disease-related label in terms of shuffle invariance, nor does it fit a parameter to the target metric and then report it as a prediction. The qualitative analysis in Section 3.5 probes device and patient labels but not disease labels, which leaves the disentanglement claim under-verified; however, under-verification is distinct from the paper's derivation reducing to its own inputs. No step in the paper exhibits the required pattern of an equation or fitted quantity being equivalent to the claimed output by construction.

Assumptions & free parameters 3 free parameters · 3 assumptions · 0 invented entities

The empirical losses depend on hand-chosen coefficients, and the design relies on the unvalidated premise that time-shuffling separates disease from background. The vCLUB bound is used as a black-box independence penalty. There are no invented physical or mathematical entities with independent falsifiable handles; the two encoders are internal network components, not separately observable entities.

free parameters (3)
  • alpha1 (Siamese loss weight) = 0.5
    Hand-chosen weight for the Siamese loss in Eq. (7); the paper states it was set empirically to facilitate convergence, with no sensitivity analysis.
  • alpha2 (vCLUB MI loss weight) = 0.02
    Hand-chosen weight for the mutual-information loss in Eq. (7), set empirically without a sensitivity study.
  • margin (Siamese loss margin) = not reported
    Threshold in Eq. (6) for the Siamese loss; its value is never given, so the exact loss landscape and behavior are not reproducible.
assumptions (3)
  • domain assumption Randomly shuffling the time axis of a spectrogram removes disease-related temporal patterns while leaving disease-irrelevant static content intact.
    This premise underpins the Siamese loss (Eq. 6) and the claim that the second encoder becomes disease-irrelevant; Section 2.1 asserts it without validation, and spectral disease cues such as wheeze tones would survive time shuffling.
  • standard math The vCLUB estimator (Eqs. 2-4) is an accurate and sufficiently tight upper bound on the mutual information between the two encoders' embeddings.
    The paper inherits the CLUB theorem from [12] and does not verify the bound's tightness for high-dimensional audio embeddings.
  • domain assumption The reconstruction loss keeps the two encoders complementary so that pushing their embeddings apart does not simply discard disease-related information.
    The shared decoder and reconstruction loss (Eq. 5) are assumed to preserve all task-relevant information even while the vCLUB loss penalizes overlap between encoders.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Disentangling Dual-Encoder Masked Autoencoder for Respiratory Sound Classification." pith.science (2026). https://pith.science/paper/TPAHFTZM

@misc{pith2026250610698,
  author       = {Pith},
  title        = {Pith review of: Disentangling Dual-Encoder Masked Autoencoder for Respiratory Sound Classification},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TPAHFTZM}},
  note         = {Machine review of arXiv:2506.10698}
}
read the original abstract

Deep neural networks have been applied to audio spectrograms for respiratory sound classification, but it remains challenging to achieve satisfactory performance due to the scarcity of available data. Moreover, domain mismatch may be introduced into the trained models as a result of the respiratory sound samples being collected from various electronic stethoscopes, patient demographics, and recording environments. To tackle this issue, we proposed a modified MaskedAutoencoder(MAE) model, named Disentangling Dual-Encoder MAE (DDE-MAE) for respiratory sound classification. Two independent encoders were designed to capture disease-related and disease-irrelevant information separately, achieving feature disentanglement to reduce the domain mismatch. Our method achieves a competitive performance on the ICBHI dataset.

Figures

Figures reproduced from arXiv: 2506.10698 by the authors.

Figure 1
Figure 1. Overview of Proposed DDE-MAE coded representations, ensuring that the encoder captures mean￾ingful features of the input data[11]. To disentangle disease-related and disease-irrelevant infor￾mation, we modify the MAE architecture to include two inde￾pendent encoders: Disease-Related Encoder. This encoder (denoted as Er) captures features directly related to respiratory diseases by pro￾cessing the input spectrograms … view at source ↗
Figure 2
Figure 2. Analysis of Encoder Performance on Device and Pa￾tient Classification 3.6. Comparison to State-of-the-art Methods Many deep neural methods have been utilized for respiratory sound classification, such as LungBRN[18], LungAttn[19], ResNeSt[20], RespireNet[21], ARSCNet[22], CNN6[23], Co-tuned ResNet-50[24], audio spectrogram transformer with patch-mix and contrastive learning (AST + patch-mix CL)[5], stethoscope-guide… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

31 extracted references · 26 canonical work pages

  1. [1]

    Accurate clas- sification can assist healthcare professionals in identifying con- ditions such as asthma, bronchitis, and pneumonia from respi- ratory sounds[1]

    Introduction Respiratory sound classification is an essential task in the early detection and diagnosis of respiratory diseases. Accurate clas- sification can assist healthcare professionals in identifying con- ditions such as asthma, bronchitis, and pneumonia from respi- ratory sounds[1]. However, several challenges arise due to the variability in record...

    work page

  2. [2]

    Methodology 2.1. Overview of Proposed DDE-MAE Our proposed model, Disentangling Dual-Encoder MAE (DDE- MAE), leverages the principles of the Masked Autoencoder (MAE) architecture to address the domain adaptation chal- lenges in respiratory sound classification. The MAE architec- ture typically consists of an encoder-decoder structure. The en- coder proces...

    work page Pith review arXiv 2025

  3. [3]

    Fine-tuning for reconstruction and classification

    Experiments 3.1. Dataset The proposed DDE-MAE is evaluated on the largest pub- licly available respiratory sound dataset, the ICBHI dataset[10], which contains 6,898 respiratory cycles, including 3,642 nor- mal breathing, 1,864 crackling breathing, 886 wheezing breath- ing, and 506 breathing with both crackling and wheezing. The sampling rates are 4kHz, 1...

    work page arXiv

  4. [4]

    Conclusion In this work, we propose the Disentangling Dual-Encoder Masked Autoencoder (DDE-MAE) for respiratory sound clas- sification, addressing domain adaptation challenges due to data variability from different patients, stethoscopes, and environ- ments. By employing two encoders—one capturing disease- related features and the other capturing disease-...

    work page

  5. [5]

    Acknowledgements Thanks to the National Natural Science Foundation of China(Grant No.62371407 and No.62001405) for funding

    work page

  6. [6]

    How the covid-19 pandemic will change the future of critical care,

    Y . M. Arabi, E. Azoulay, H. M. Al-Dorzi, J. Phua, J. Salluh, A. Binnie, C. Hodgson, D. C. Angus, M. Cecconi, B. Duet al., “How the covid-19 pandemic will change the future of critical care,”Intensive care medicine, vol. 47, pp. 282–291, 2021

    work page 2021

  7. [7]

    Domain adaptation for medical image anal- ysis: a survey,

    H. Guan and M. Liu, “Domain adaptation for medical image anal- ysis: a survey,”IEEE Transactions on Biomedical Engineering, vol. 69, no. 3, pp. 1173–1185, 2021

    work page 2021

  8. [8]

    Qui ˜nonero-Candela, M

    J. Qui ˜nonero-Candela, M. Sugiyama, A. Schwaighofer, and N. D. Lawrence,Dataset shift in machine learning. Mit Press, 2022

    work page 2022

Show all 31 references
  1. [9]

    Unbiased look at dataset bias,

    A. Torralba and A. A. Efros, “Unbiased look at dataset bias,” in CVPR 2011. IEEE, 2011, pp. 1521–1528

    2011

  2. [10]

    Patch-mix contrastive learning with audio spectrogram transformer on respiratory sound classifi- cation,

    S. Bae, J.-W. Kim, W.-Y . Cho, H. Baek, S. Son, B. Lee, C. Ha, K. Tae, S. Kim, and S.-Y . Yun, “Patch-mix contrastive learning with audio spectrogram transformer on respiratory sound classifi- cation,” inINTERSPEECH 2023, 2023, pp. 5436–5440

    2023

  3. [11]

    Multi-view spectro- gram transformer for respiratory sound classification,

    W. He, Y . Yan, J. Ren, R. Bai, and X. Jiang, “Multi-view spectro- gram transformer for respiratory sound classification,” inICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2024, pp. 8626–8630

    2024

  4. [12]

    Bts: Bridging text and sound modalities for metadata-aided res- piratory sound classification,

    J.-W. Kim, M. Toikkanen, Y . Choi, S.-E. Moon, and H.-Y . Jung, “Bts: Bridging text and sound modalities for metadata-aided res- piratory sound classification,”arXiv preprint arXiv:2406.06786, 2024

    2024 arXiv

  5. [13]

    Domain-adversarial training of neural networks,

    Y . Ganin, E. Ustinova, H. Ajakan, P. Germain, H. Larochelle, F. Laviolette, M. March, and V . Lempitsky, “Domain-adversarial training of neural networks,”Journal of machine learning re- search, vol. 17, no. 59, pp. 1–35, 2016

    2016

  6. [14]

    Stethoscope-guided supervised contrastive learning for cross- domain adaptation on respiratory sound classification,

    J.-W. Kim, S. Bae, W.-Y . Cho, B. Lee, and H.-Y . Jung, “Stethoscope-guided supervised contrastive learning for cross- domain adaptation on respiratory sound classification,” inICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). ...

    2024

  7. [15]

    A respiratory sound database for the development of auto- mated classification,

    B. Rocha, D. Filos, L. Mendes, I. V ogiatzis, E. Perantoni, E. Kaimakamis, P. Natsiavas, A. Oliveira, C. J´acome, A. Marques et al., “A respiratory sound database for the development of auto- mated classification,” inPrecision Medicine Powered by pHealth and Connected Health: ...

    2017

  8. [16]

    Masked autoencoders are scalable vision learners,

    K. He, X. Chen, S. Xie, Y . Li, P. Doll´ar, and R. Girshick, “Masked autoencoders are scalable vision learners,” inCVPR, 2022, pp. 16 000–16 009

    2022

  9. [17]

    CLUB: A contrastive log-ratio upper bound of mutual information,

    P. Cheng, W. Hao, S. Dai, J. Liu, Z. Gan, and L. Carin, “CLUB: A contrastive log-ratio upper bound of mutual information,” in Proceedings of the 37th International Conference on Machine Learning, ser. Proceedings of Machine Learning Research, H. D. III and A. Singh, Eds., vol....

    2020

  10. [18]

    Siamese neural networks: An overview,

    D. Chicco, “Siamese neural networks: An overview,”Artificial neural networks, pp. 73–94, 2021

    2021

  11. [19]

    A mathematical theory of communication,

    C. E. Shannon, “A mathematical theory of communication,”The Bell system technical journal, vol. 27, no. 3, pp. 379–423, 1948

    1948

  12. [20]

    Masked autoencoders that lis- ten,

    P.-Y . Huang, H. Xu, J. Li, A. Baevski, M. Auli, W. Galuba, F. Metze, and C. Feichtenhofer, “Masked autoencoders that lis- ten,”Advances in Neural Information Processing Systems, vol. 35, pp. 28 708–28 720, 2022

    2022

  13. [21]

    Audio set: An ontology and human-labeled dataset for audio events,

    J. F. Gemmeke, D. P. Ellis, D. Freedman, A. Jansen, W. Lawrence, R. C. Moore, M. Plakal, and M. Ritter, “Audio set: An ontology and human-labeled dataset for audio events,” in2017 IEEE inter- national conference on acoustics, speech and signal processing (ICASSP). IEEE, 2017, ...

    2017

  14. [22]

    Decoupled weight de- cay regularization,

    I. Loshchilov and F. Hutter, “Decoupled weight de- cay regularization,” inInternational Conference on Learn- ing Representations, 2017. [Online]. Available: https: //api.semanticscholar.org/CorpusID:53592270

    2017

  15. [23]

    Lungbrn: A smart digital stethoscope for detecting respiratory disease using bi-resnet deep learning algorithm,

    Y . Ma, X. Xu, Q. Yu, Y . Zhang, Y . Li, J. Zhao, and G. Wang, “Lungbrn: A smart digital stethoscope for detecting respiratory disease using bi-resnet deep learning algorithm,” in2019 IEEE Biomedical Circuits and Systems Conference (BioCAS). IEEE, 2019, pp. 1–4

    2019

  16. [24]

    Lungattn: advanced lung sound classification using attention mechanism with dual tqwt and triple stft spectrogram,

    J. Li, J. Yuan, H. Wang, S. Liu, Q. Guo, Y . Ma, Y . Li, L. Zhao, and G. Wang, “Lungattn: advanced lung sound classification using attention mechanism with dual tqwt and triple stft spectrogram,” Physiological Measurement, vol. 42, no. 10, p. 105006, 2021

    2021

  17. [25]

    A domain transfer based data augmenta- tion method for automated respiratory classification,

    Z. Wang and Z. Wang, “A domain transfer based data augmenta- tion method for automated respiratory classification,” inICASSP 2022-2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2022, pp. 9017–9021

    2022

  18. [26]

    Respirenet: A deep neural network for accurately detecting abnormal lung sounds in limited data setting,

    S. Gairola, F. Tom, N. Kwatra, and M. Jain, “Respirenet: A deep neural network for accurately detecting abnormal lung sounds in limited data setting,” in2021 43rd Annual International Confer- ence of the IEEE Engineering in Medicine & Biology Society (EMBC). IEEE, 2021, pp. 527–530

    2021

  19. [27]

    Arsc- net: Adventitious respiratory sound classification network using parallel paths with channel-spatial attention,

    L. Xu, J. Cheng, J. Liu, H. Kuang, F. Wu, and J. Wang, “Arsc- net: Adventitious respiratory sound classification network using parallel paths with channel-spatial attention,” in2021 IEEE Inter- national Conference on Bioinformatics and Biomedicine (BIBM). IEEE, 2021, pp. 1125–1130

    2021

  20. [28]

    Pretraining respiratory sound rep- resentations using metadata and contrastive learning,

    I. Moummad and N. Farrugia, “Pretraining respiratory sound rep- resentations using metadata and contrastive learning,” in2023 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA). IEEE, 2023, pp. 1–5

    2023

  21. [29]

    Lung sound classification using co-tuning and stochastic normalization,

    T. Nguyen and F. Pernkopf, “Lung sound classification using co-tuning and stochastic normalization,”IEEE Transactions on Biomedical Engineering, vol. 69, no. 9, pp. 2872–2882, 2022

    2022

  22. [30]

    mixup: Beyond empirical risk minimization,

    H. Zhang, M. Ciss ´e, Y . N. Dauphin, and D. Lopez-Paz, “mixup: Beyond empirical risk minimization,” in6th International Conference on Learning Representations, ICLR 2018, Vancouver, BC, Canada, April 30 - May 3, 2018, Conference Track Proceedings. OpenReview.net, 2018. [Onlin...

    2018

  23. [31]

    Cutmix: Regularization strategy to train strong classifiers with localizable features,

    S. Yun, D. Han, S. J. Oh, S. Chun, J. Choe, and Y . Yoo, “Cutmix: Regularization strategy to train strong classifiers with localizable features,” inProceedings of the IEEE/CVF international confer- ence on computer vision, 2019, pp. 6023–6032

    2019

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.