Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Cough Classification using Few-Shot Learning

T0 review · 4 major / 5 minor · reviewed 2026-08-04 · deepseek-v4-flash

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

Pith's one-line read A few-shot learning model using Prototypical Networks can classify COVID-19, flu, and healthy coughs from as few as 15 labeled examples per class, achieving accuracy comparable to binary classifiers within a ±15 percentage-point equivalence

desk verdict A reproducible few-shot cough study whose central disease-discrimination claim is undermined by perfect confounding between class label and recording source; the binary-vs-multiclass comparison survives, the absolute accuracies do not. read the letter →

arxiv 2509.09515 v1 pith:6CM6GRF5 submitted 2025-09-11 cs.LG

Yoga Disha Sendhil Kumar , Manas V Shetty , Sudip Vhaduri This is my paper · ORCID
classification cs.LG
keywords few-shotlearningcoughclassificationPrototypicalNetworksMel-spectrogramCOVID-19detectioninfluenzaepisodictrainingrespiratorysound
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 argues that few-shot learning can make cough-based disease classification practical when labeled data is scarce. Using Prototypical Networks over Mel-spectrograms, the authors train a three-way COVID-19/flu/healthy classifier with only 15 support examples per class and report accuracy around 72–75%. They further claim this single multi-class model is statistically equivalent to the three binary classifiers within a preset ±15 percentage-point margin, based on equivalence tests. If true, a single compact model could replace multiple binary detectors in low-resource screening settings.

What carries the argument

Prototypical Networks: a few-shot method that embeds support examples into a vector space using a shared convolutional backbone, averages the embeddings per class to form a prototype, and classifies a query by its Euclidean distance to the closest prototype. The backbone is a standard image-classification network modified for single-channel spectrograms and initialized with transfer-learning weights; training is episodic, sampling N-way, K-shot tasks to mimic scarce-data conditions.

What would settle it

Run the same few-shot pipeline on a dataset where healthy, COVID-19, and flu coughs are recorded with an identical device and protocol, or swap the class labels across the three source datasets; if accuracy falls to near chance or the class ranking changes, the model has learned dataset provenance, not disease acoustics.

Watch

Extended reading notes

Core claim

The paper's central claim is that Prototypical Networks, a metric-based few-shot learning method, can classify cough sounds into three conditions—healthy, COVID-19, and flu—with accuracy competitive with traditional deep learning models trained on far more data. With 15 labeled support examples per class, the three-way model reaches roughly 72% accuracy (74.87% is stated in the abstract), while binary pairs exceed 70% accuracy for all combinations. Equivalence testing with a ±15 percentage-point margin finds no significant difference between the multi-class and binary models, supporting the practical use of a single multi-class classifier. Class-wise, flu is the most distinguishable and heal

Load-bearing premise

The load-bearing premise is that the class labels reflect the disease conditions rather than the recording source; because each class is drawn from a different public dataset, differences in microphone hardware, environment, and collection protocol may be doing the work that the model credits to the disease.

Editorial extensions

If this is right

  • Deploying a three-way few-shot cough classifier is viable: it cuts the need for large labeled datasets and avoids running three separate binary models.
  • Health screening for emerging respiratory diseases could start with a small labeled set (15 per class) and still reach usable accuracy.
  • Flu coughs are more acoustically distinct, while healthy coughs need more diverse training data to be recognized reliably.
  • Accuracy gains from adding support examples diminish after 10 shots, so collecting beyond that yields little improvement.

Reading between the lines

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

  • Because each class comes from a different source dataset, part of the reported separation may be recording environment rather than disease biology; a cross-dataset or matched-recording test would settle this.
  • The ±15 percentage-point equivalence margin is wide; a tighter margin (e.g., ±5) would be needed before replacing binary models in high-stakes clinical triage.
  • The same prototypical-network pipeline could transfer to other scarce-data audio diagnostics, such as tuberculosis or asthma cough screening, but would need the same provenance controls.
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

4 major / 5 minor

Summary. The paper applies Prototypical Networks with a ResNet-18 backbone to Mel-spectrogram representations of cough sounds for three-way classification (Healthy, COVID-19, Flu). Using 100 samples per class drawn from Coswara, COUGHVID, and FluSense, the authors report binary and multi-class accuracies across K-shot settings, perform TOST and bootstrap-based equivalence tests to argue that multi-class and binary models are statistically equivalent within a ±15 percentage point margin, and compare their few-shot results qualitatively with large-data deep learning baselines. The central claims are that few-shot models can achieve competitive accuracy with very few labeled examples and that a 3-way classifier can replace three binary classifiers.

Significance. If the claims were supported, the paper would provide a useful demonstration that few-shot learning can reduce labeled-data requirements for cough-based respiratory screening, and that a single multi-class model can substitute for multiple binary models in low-resource settings. The use of an equivalence framework rather than a mere difference test is also a thoughtful choice. However, the experimental design contains a source-provenance confound that directly affects the main disease-classification claim, and several reported numbers are internally inconsistent. The significance of the contribution currently rests on these questionable foundations.

major comments (4)
  1. [Section III.A, Table III, Fig. 8] The three classes are perfectly confounded with dataset provenance: Healthy samples come only from Coswara, COVID-19 only from COUGHVID, and Flu only from FluSense. The model may therefore be learning dataset-, microphone-, or environment-specific artifacts rather than cough pathology. This threatens the absolute accuracy claims, the class-level conclusion that Flu is most distinguishable (96.8% at K=15, Table III), and the t-SNE interpretation in Fig. 8. The Discussion acknowledges self-reported labels and healthy-class ambiguity but does not address this confound. A within-source or cross-dataset control (e.g., using Coswara for both healthy and COVID, or a held-out dataset for each class) is needed to support the disease-classification claim.
  2. [Abstract; Section IV.A, Tables I-II; Section IV.C] Several reported numbers are inconsistent. The Abstract states 74.87% multi-class accuracy at 15 support examples, but Table II reports 72.07±2.31% at K=15. Table IV reports a multi-class mean of 73.22%, whereas the average of the Table II values (60.67, 68.53, 71.93, 72.07) is 68.3%. The Abstract also quotes paired t-test p=0.149 and Wilcoxon p=0.125, but Section IV.C presents only TOST/bootstrap equivalence results and does not report these p-values. These discrepancies prevent reproducibility and must be resolved.
  3. [Section IV.C] The equivalence margin of ±15 percentage points is very wide and is not justified by any diagnostic standard. The framing in the Introduction (Section I.B) mentions a 10–15% accuracy margin, but the tests use 15%, which is the most permissive end. Aggregating over all K-shot settings (400 multi-class episodes vs. 1200 binary episodes) mixes regimes with very different difficulty; an equivalence claim should be demonstrated at each K or with a prespecified and clinically motivated margin. With such a wide margin, the claim that the two model types are 'statistically equivalent' is weak.
  4. [Section IV.E] The comparison to traditional deep learning approaches is narrative and not a controlled comparison. Prior studies are cited with different datasets, preprocessing, class definitions, and evaluation protocols; no baseline model is trained on the same curated 100-per-class dataset. The claim that few-shot models 'achieve competitive accuracy' with large-data approaches is therefore not substantiated. A same-data baseline (e.g., a standard supervised CNN trained on the full training set) should be included.
minor comments (5)
  1. [References] References [5] and [17] are duplicates (COUGHVID), [6] and [16] are duplicates (Coswara), and [7] and [18] are duplicates (FluSense). Please consolidate.
  2. [Fig. 7] The caption says 'Confusion Matrix for 100 Few-Shot Classification Tasks' but the figure's content is not described in the text, and the figure is not referenced in the body. Please either describe it or remove it.
  3. [Fig. 8] Typo in caption: 'Data Distribution In-terms' should be 'Data Distribution in Terms'.
  4. [Section III.D] The EasyFSL framework is not described or cited, and no version/link is given. Please provide details or a reference.
  5. [Section III.D and IV.A] Training hyperparameters are only described as 'tune learning rates'; the actual optimizer settings, learning rate, and number of epochs are not reported. The average over 100 episodes is stated, but the number of query examples per episode is not given in the tables or text. Please provide these details.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the reported accuracy and equivalence results are empirical outcomes of a train/evaluate pipeline, not consequences of a fitted input or self-citation chain.

full rationale

The paper's central claims are empirical. The multi-class and binary accuracies (Tables I--II) are obtained by training prototypical networks on episodic support/query splits and averaging over 100 episodes per K-shot setting; no accuracy value is derived from a parameter fitted to the test set. The TOST and bootstrap equivalence tests use a pre-specified ±15% equivalence margin and compute a 90% confidence interval that lies inside it; the margin is an evaluation threshold, not an output of the fitting procedure. The architecture components (Prototypical Networks, ResNet-18, Mel-spectrograms) are standard and cited to external sources (Snell et al. [3]; He et al. [19]), not to a self-citation chain. The many self-citations ([20]--[44]) are contextual references to the authors' broader smart-health work and do not support any load-bearing derivation. The principal weakness is a dataset confound: each class is drawn from a different source (Coswara for Healthy, COUGHVID for COVID-19, FluSense for Flu), so class-level accuracies and the t-SNE separation may reflect recording provenance rather than disease acoustics. This is a validity/correctness threat, not a circularity threat: the model is not defined in terms of its own outputs, and the comparison to external large-data baselines, while not quantitative, does not reduce to the paper's own fitted values. No step in the paper's derivation chain is equivalent to its input by construction. Therefore the circularity score is 0.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The central evaluation rests on standard few-shot machinery plus several domain assumptions. No invented entities are introduced. The main hand-chosen number is the 15 point equivalence margin, which defines the paper's central conclusion.

free parameters (2)
  • Equivalence margin = plus or minus 15 percentage points
    Hand-chosen threshold in Section IV.C; the TOST/bootstrap conclusion of statistical equivalence holds only within this wide margin.
  • Training hyperparameters (learning rate, epochs) = not reported
    Section III.D says learning rates are tuned and training took about 6 hours, but no values or seeds are given; these choices affect the reported accuracies.
assumptions (5)
  • standard math Prototypical Networks with Euclidean distance in embedding space provide a valid classifier for few-shot tasks
    Adopted from Snell et al. [3]; the paper does not re-derive it.
  • domain assumption Mel-spectrograms computed with Librosa preserve the acoustic differences between healthy, COVID-19, and flu coughs
    Section III.B uses 128-band mel-spectrograms as model input without validation against other features.
  • domain assumption Self-reported labels in Coswara, COUGHVID, and FluSense are sufficiently accurate for supervised training
    Section III.A and Section V acknowledge label noise and self-report ambiguity.
  • domain assumption Each class can be represented by one dataset, so cough differences are not confounded by recording conditions
    Section III.A assigns Healthy to Coswara, COVID-19 to COUGHVID, and Flu to FluSense; this is not controlled.
  • ad hoc to paper An equivalence margin of plus or minus 15 percentage points is the correct standard for comparable diagnostic performance
    Section IV.C sets this margin; with a smaller margin the equivalence conclusion would likely fail.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Cough Classification using Few-Shot Learning." pith.science (2026). https://pith.science/paper/6CM6GRF5

@misc{pith2026250909515,
  author       = {Pith},
  title        = {Pith review of: Cough Classification using Few-Shot Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6CM6GRF5}},
  note         = {Machine review of arXiv:2509.09515}
}
read the original abstract

This paper investigates the effectiveness of few-shot learning for respiratory sound classification, focusing on coughbased detection of COVID-19, Flu, and healthy conditions. We leverage Prototypical Networks with spectrogram representations of cough sounds to address the challenge of limited labeled data. Our study evaluates whether few-shot learning can enable models to achieve performance comparable to traditional deep learning approaches while using significantly fewer training samples. Additionally, we compare multi-class and binary classification models to assess whether multi-class models can perform comparably to their binary counterparts. Experimental findings show that few-shot learning models can achieve competitive accuracy. Our model attains 74.87% accuracy in multi-class classification with only 15 support examples per class, while binary classification achieves over 70% accuracy across all class pairs. Class-wise analysis reveals Flu as the most distinguishable class, and Healthy as the most challenging. Statistical tests (paired t-test p = 0.149, Wilcoxon p = 0.125) indicate no significant performance difference between binary and multiclass models, supporting the viability of multi-class classification in this setting. These results highlight the feasibility of applying few-shot learning in medical diagnostics, particularly when large labeled datasets are unavailable.

Figures

Figures reproduced from arXiv: 2509.09515 by the authors.

Figure 1
Figure 1. Visualization of support set spectrograms used for [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Visualization of query set spectrograms used for [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 4
Figure 4. Binary Classification Accuracy COVID-19 vs. Healthy remains the most challenging pair, reaching 74.80% at K = 15. Figures 4 and 5 illustrate these trends. In [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗
Figures from the paper (4 more)
Figure 5
Figure 5. Figure 5: Multi-Class Accuracy [PITH_FULL_IMAGE:figures/full_fig_p004_5.png]
Figure 6
Figure 6. Figure 6: Class-level Average Accuracy (%) of Multi-Class [PITH_FULL_IMAGE:figures/full_fig_p005_6.png]
Figure 7
Figure 7. Figure 7: Confusion Matrix for 100 Few-Shot Classification [PITH_FULL_IMAGE:figures/full_fig_p005_7.png]
Figure 8
Figure 8. Figure 8: t-SNE Visualization of Data Distribution In-terms of [PITH_FULL_IMAGE:figures/full_fig_p006_8.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. HEARTS: Benchmarking LLM Reasoning on Health Time Series

    cs.LG 2026-02 conditional novelty 7.0 of 10

    A 110-task benchmark across 20 health signal modalities shows current LLMs underperform specialized models and depend on simple heuristics rather than robust time-series reasoning.

Reference graph

Works this paper leans on

46 extracted references · 4 linked inside Pith · cited by 1 Pith paper

  1. [1]

    Imran, J

    M. Imran, J. Posokhova, A. Qureshi, et al., ”AI4COVID-19: AI enabled preliminary diagnosis for COVID-19 from cough samples via an app,” Informatics in Medicine Unlocked, vol. 20, 2020

    2020

  2. [2]

    Brownlee, ”Deep Learning for Time Series Forecasting: Predict the Future with MLPs, CNNs and LSTMs in Python,”Machine Learning Mastery, 2018

    J. Brownlee, ”Deep Learning for Time Series Forecasting: Predict the Future with MLPs, CNNs and LSTMs in Python,”Machine Learning Mastery, 2018

    2018

  3. [3]

    Snell, K

    J. Snell, K. Swersky, and R. Zemel, ”Prototypical Networks for Few- shot Learning,”Advances in Neural Information Processing Systems (NeurIPS), 2017

    2017

  4. [4]

    Laguarta, F

    A. Laguarta, F. Hueto, and B. Subirana, ”COVID-19 artificial intelli- gence diagnosis using only cough recordings,”IEEE Open Journal of Engineering in Medicine and Biology, vol. 1, pp. 275-281, 2020

    2020

  5. [8]

    C. Brown et al., ”Exploring automatic diagnosis of COVID-19 from crowdsourced respiratory sound data,”Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, 2020

    2020

  6. [9]

    Deshpande et al., ”A review of deep learning techniques for COVID- 19 detection using radiological images,”Artificial Intelligence Review, vol

    A. Deshpande et al., ”A review of deep learning techniques for COVID- 19 detection using radiological images,”Artificial Intelligence Review, vol. 54, no. 1, pp. 457-497, 2021

    2021

  7. [10]

    Nanni, S

    T. Nanni, S. Ghidoni, and S. Brahnam, ”Handcrafted vs. non-handcrafted features for computer vision classification,”Pattern Recognition, vol. 71, pp. 158-172, 2017

    2017

  8. [11]

    Antoniou, H

    A. Antoniou, H. Edwards, and A. Storkey, ”How to train your MAML,” International Conference on Learning Representations (ICLR), 2019

    2019

Show all 46 references
  1. [12]

    H. Shibata et al., ”Few-shot learning for sound event classification via meta-learning,”Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2021

    2021

  2. [13]

    Y . Wang, R. Zhang, and P. Liang, ”Few-shot learning for respiratory sound classification,”Proceedings of the 2022 IEEE International Con- ference on Acoustics, Speech, and Signal Processing (ICASSP), 2022

    2022

  3. [14]

    Chen et al., ”A new episodic training method for few-shot learning,” Neural Networks, vol

    X. Chen et al., ”A new episodic training method for few-shot learning,” Neural Networks, vol. 144, pp. 117-127, 2021

    2021

  4. [15]

    Ravi and A

    M. Ravi and A. Larochelle, ”Optimization as a model for few-shot learn- ing,”International Conference on Learning Representations (ICLR), 2017

    2017

  5. [16]

    Sharma et al., ”Coswara–A database of breathing, cough, and voice sounds for COVID-19 diagnosis,”arXiv preprint arXiv:2005.10548, 2020

    P. Sharma et al., ”Coswara–A database of breathing, cough, and voice sounds for COVID-19 diagnosis,”arXiv preprint arXiv:2005.10548, 2020

    2005 arXiv

  6. [17]

    Orlandic, T

    R. Orlandic, T. Teijeiro, and D. Atienza, ”The COUGHVID crowd- sourcing dataset: A corpus for the study of large-scale cough analysis algorithms,”Scientific Data, vol. 8, 2021

    2021

  7. [18]

    M. Rahman et al., ”FluSense: A contactless syndromic surveillance platform for inFluenza-like illness in public spaces,”Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies, vol. 4, no. 1, pp. 1-28, 2020

    2020

  8. [19]

    K. He, X. Zhang, S. Ren, and J. Sun, ”Deep residual learning for image recognition,”Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016

    2016

  9. [20]

    Transfer learning to detect covid-19 coughs with incremental addition of patient coughs to healthy people’s cough detection models,

    S. Vhaduri, S. Paik, and J.E. Huber, “Transfer learning to detect covid-19 coughs with incremental addition of patient coughs to healthy people’s cough detection models,” International Conference on Wireless Mobile Communication and Healthcare, 2023

    2023

  10. [21]

    Predicting a user’s demo- graphic identity from leaked samples of health-tracking wearables and understanding associated risks,

    S. Vhaduri, S.V . Dibbo, and C.-Y . Chen, “Predicting a user’s demo- graphic identity from leaked samples of health-tracking wearables and understanding associated risks,” 2022 IEEE 10th International Confer- ence on Healthcare Informatics (ICHI), pp. 309–318, 2022

    2022

  11. [22]

    Bag of on-phone ANNs to secure IoT objects using wearable and smartphone biometrics,

    S. Vhaduri, W. Cheung, and S.V . Dibbo, “Bag of on-phone ANNs to secure IoT objects using wearable and smartphone biometrics,” IEEE Transactions on Dependable and Secure Computing, 2023

    2023

  12. [23]

    mWIoTAuth: Multi-wearable data-driven implicit IoT authentication,

    S. Vhaduri, S.V . Dibbo, A. Muratyan, and W. Cheung, “mWIoTAuth: Multi-wearable data-driven implicit IoT authentication,” Future Gener- ation Computer Systems, vol. 159, pp. 230–242, 2024

    2024

  13. [24]

    Implicit IoT authentication using on-phone ANN models and breathing data,

    S. Vhaduri, S.V . Dibbo, and W. Cheung, “Implicit IoT authentication using on-phone ANN models and breathing data,” Internet of Things, vol. 24, p. 101003, 2023

    2023

  14. [25]

    Towards reliable wearable-user identifi- cation,

    S. Vhaduri and C. Poellabauer, “Towards reliable wearable-user identifi- cation,” 2017 IEEE International Conference on Healthcare Informatics (ICHI), pp. 329–329, 2017

    2017

  15. [26]

    Biometric-based wearable user authen- tication during sedentary and non-sedentary periods,

    S. Vhaduri and C. Poellabauer, “Biometric-based wearable user authen- tication during sedentary and non-sedentary periods,” arXiv preprint arXiv:1811.07060, 2018

    2018 arXiv

  16. [27]

    On-phone CNN model-based implicit authentication to secure IoT wearables,

    S.V . Dibbo, W. Cheung, and S. Vhaduri, “On-phone CNN model-based implicit authentication to secure IoT wearables,” The Fifth International Conference on Safety and Security with IoT: SaSeIoT 2021, pp. 19–34, 2022

    2021

  17. [28]

    Continuous Authentication of Wearable Device Users from Heart Rate, Gait, and Breathing Data,

    W. Cheung and S. Vhaduri, “Continuous Authentication of Wearable Device Users from Heart Rate, Gait, and Breathing Data,” IEEE RAS & EMBS International Conference on Biomedical Robotics and Biomecha- tronics (BioRob), 2020

    2020

  18. [29]

    Context-Dependent Implicit Authentica- tion for Wearable Device Users,

    W. Cheung and S. Vhaduri, “Context-Dependent Implicit Authentica- tion for Wearable Device Users,” IEEE International Symposium on Personal, Indoor, and Mobile Radio Communications (PIMRC), 2020

    2020

  19. [30]

    Understanding user concerns and choice of app architectures in designing audio-based mHealth apps,

    S. Vhaduri and S.S. Simhadri, “Understanding user concerns and choice of app architectures in designing audio-based mHealth apps,” Smart Health, vol. 26, p. 100341, 2022

    2022

  20. [31]

    Design and Implementation of a Remotely Configurable and Manageable Well-being Study,

    S. Vhaduri and C. Poellabauer, “Design and Implementation of a Remotely Configurable and Manageable Well-being Study,” EAI SWIT- Health, 2015

    2015

  21. [32]

    Estimating sleep duration from temporal factors, daily activities, and smartphone use,

    C.-Y . Chen, S. Vhaduri, and C. Poellabauer, “Estimating sleep duration from temporal factors, daily activities, and smartphone use,” 2020 IEEE 44th Annual Computers, Software, and Applications Conference (COMPSAC), pp. 545–554, 2020

    2020

  22. [33]

    Nocturnal cough and snore detection using smartphones in presence of multiple background-noises,

    S. Vhaduri, “Nocturnal cough and snore detection using smartphones in presence of multiple background-noises,” Proceedings of the 3rd ACM SIGCAS Conference on Computing and Sustainable Societies, pp. 174– 186, 2020

    2020

  23. [34]

    Adherence to personal health devices: A case study in diabetes management,

    S. Vhaduri and T. Prioleau, “Adherence to personal health devices: A case study in diabetes management,” Proceedings of the 14th EAI International Conference on Pervasive Computing Technologies for Healthcare, pp. 62–72, 2020

    2020

  24. [35]

    Effect of noise on generic cough models,

    S.V . Dibbo, Y . Kim, and S. Vhaduri, “Effect of noise on generic cough models,” 2021 IEEE 17th International Conference on Wearable and Implantable Body Sensor Networks (BSN), pp. 1–4, 2021

    2021

  25. [36]

    Environment knowledge-driven generic models to detect coughs from audio recordings,

    S. Vhaduri, S.V . Dibbo, and Y . Kim, “Environment knowledge-driven generic models to detect coughs from audio recordings,” IEEE Open Journal of Engineering in Medicine and Biology, vol. 4, pp. 55–66, 2023

    2023

  26. [37]

    Deriving college students’ phone call patterns to improve student life,

    S. Vhaduri, S.V . Dibbo, and Y . Kim, “Deriving college students’ phone call patterns to improve student life,” IEEE Access, vol. 9, pp. 96453– 96465, 2021

    2021

  27. [38]

    Visualizing college students’ geo-temporal context-varying significant phone call patterns,

    S.V . Dibbo, Y . Kim, S. Vhaduri, and C. Poellabauer, “Visualizing college students’ geo-temporal context-varying significant phone call patterns,” 2021 IEEE 9th International Conference on Healthcare Informatics (ICHI), pp. 381–385, 2021

    2021

  28. [39]

    Predicting next call duration: A future direction to promote mental health in the age of lockdown,

    S. Vhaduri, S.V . Dibbo, C.-Y . Chen, and C. Poellabauer, “Predicting next call duration: A future direction to promote mental health in the age of lockdown,” 2021 IEEE 45th Annual Computers, Software, and Applications Conference (COMPSAC), pp. 804–811, 2021

    2021

  29. [40]

    Understanding college stu- dents’ phone call behaviors towards a sustainable mobile health and wellbeing solution,

    Y . Kim, S. Vhaduri, and C. Poellabauer, “Understanding college stu- dents’ phone call behaviors towards a sustainable mobile health and wellbeing solution,” arXiv preprint arXiv:2011.06007, 2020

    2011 arXiv

  30. [41]

    Opportunistic discovery of personal places using multi-source sensor data,

    S. Vhaduri and C. Poellabauer, “Opportunistic discovery of personal places using multi-source sensor data,” IEEE Transactions on Big Data, vol. 7, no. 2, pp. 383–396, 2018

    2018

  31. [42]

    Opportunistic discovery of personal places using smartphone and fitness tracker data,

    S. Vhaduri and C. Poellabauer, “Opportunistic discovery of personal places using smartphone and fitness tracker data,” 2018 IEEE Interna- tional Conference on Healthcare Informatics (ICHI), pp. 103–114, 2018

    2018

  32. [43]

    Audio-Based Clas- sification of Insect Species Using Machine Learning Models,

    Yoga Disha Sendhil Kumar, and Manas V Shetty, “Audio-Based Clas- sification of Insect Species Using Machine Learning Models,”arXiv preprint arXiv:2502.13893, 2024. Available: https://arxiv.org/abs/2502. 13893

    2024 arXiv

  33. [44]

    Facial recognition to detect mood and play songs,

    Yogadisha Sendhil Kumar, Sumkh R. R, Manas V shetty and Rahul Reddy, “Facial recognition to detect mood and play songs,” inProceed- ings of the IEMIS 2022, LNNS, vol. 592

    2022

  34. [45]

    Few-shot learning in medical imaging: A systematic review,

    J.S. Suri, R.M. Laxminarayan, and K. Dutta, “Few-shot learning in medical imaging: A systematic review,”Computers in Biology and Medicine, vol. 170, 107618, 2024

    2024

  35. [46]

    Transfer learning in medical image classification: A comprehensive review,

    A. Islam, A. Jahan, and M.A. Hossain, “Transfer learning in medical image classification: A comprehensive review,”Sustainability, vol. 15, no. 7, pp. 5930, 2023

    2023

  36. [47]

    Multistage transfer learning for improving deep learning performance in medical imaging,

    Y . Yan, X. Tang, and W. Zuo, “Multistage transfer learning for improving deep learning performance in medical imaging,”Artificial Intelligence Review, 2024

    2024

  37. [48]

    An improved few-shot learning method for tumor image recognition,

    Y . Zhang, H. Li, and M. Zhang, “An improved few-shot learning method for tumor image recognition,”Frontiers in Computational Neuroscience, vol. 17, 2023

    2023

  38. [49]

    Transfer learning-based COVID- 19 detection from chest X-ray images with reduced training samples,

    M.M. Rahman, S. Islam, and S. Satu, “Transfer learning-based COVID- 19 detection from chest X-ray images with reduced training samples,” BMC Medical Imaging, vol. 22, no. 1, pp. 115, 2022

    2022

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.