Pith. sign in

REVIEW 3 major objections 5 minor 96 references

Versatile Cardiovascular Signal Generation with a Unified Diffusion Transformer

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 ↗

Pith's one-line read One diffusion transformer can denoise, fill in, and translate PPG, ECG, and blood-pressure signals.

desk verdict A genuinely useful unified PPG/ECG/BP generation framework, but the headline claim overstates and the central 'unified time step' mechanism isn't what Algorithm 1 implements. read the letter →

arxiv 2505.22306 v2 pith:YMISX4OE submitted 2025-05-28 cs.LG cs.AI

Zehua Chen , Yuyang Miao , Liyuan Wang , Luyun Fan , Danilo P. Mandic , Jun Zhu This is my paper
classification cs.LGcs.AI
keywords diffusionmodeltransformercardiovascularsignalsphotoplethysmography(PPG)electrocardiography(ECG)bloodpressuremulti-modalgenerationcontinuallearning
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

The paper tries to establish that a single diffusion transformer, UniCardio, can handle the three main cardiovascular signal modalities—PPG, ECG, and blood pressure—within one generative framework, covering denoising, imputation, and cross-modal translation. The motivation is that these signals are complementary views of the same cardiovascular process, so a model that learns their joint conditional distributions should restore corrupted recordings and synthesize unrecorded signals better than task-specific methods. The paper reports that UniCardio outperforms recent task-specific baselines on denoising, imputation, and translation benchmarks, and that its generated signals support downstream detection of abnormalities and vital-sign estimation at a level comparable to ground truth. If true, this would mean a single compact model could replace a collection of special-purpose signal-processing systems in real-time cardiovascular monitoring.

What carries the argument

The load-bearing mechanism is a conditional diffusion transformer with three coordinated pieces: modality-specific multi-scale convolutional encoders that map each signal into a shared latent timeline; customized transformer modules with task-specific attention masks that zero out interactions except those from condition to target modalities; and a phased continual learning paradigm that trains on one-, two-, and three-condition tasks with learning-rate scheduling, replay of earlier task batches, and masked attention to prevent catastrophic forgetting. A distinctive simplifying choice is the unified diffusion time step: instead of giving each modality its own noise level at each step, UniCardio applies one shared time step $t$ to all condition and target signals, so the single noise estimator $\epsilon_\theta(x_t, c, t)$ must represent every inter-modal conditional distribution.

What would settle it

Train a variant of UniCardio that uses per-modality diffusion time steps (one noise level for each condition modality and one for the target) on the same Cuffless BP tasks; if it beats the unified-time-step model on PPG-to-ECG or PPG-to-BP RMSE, the shared-schedule assumption is the limiting factor. A complementary test is to record PPG from a wrist wearable during motion and compare whether UniCardio's denoised and translated ECG still restores AF and ST-change detector performance to ground-truth levels.

Watch

Extended reading notes

Core claim

UniCardio's central claim is that arbitrary many-to-any conditional distributions among PPG, ECG, and BP can be learned by one noise-prediction network, using an unconditional forward process that maps all modalities to a shared Gaussian prior and a conditional reverse process guided by observed modalities. The model couples this with task-specific attention masks that allow only condition-to-target token interactions during self-attention, and a continual learning schedule that introduces tasks with one, two, and three condition modalities in phases. The paper reports that this unified model achieves lower RMSE and better distributional fidelity than task-specific baselines such as PulseImpute, DeepMVI, RDDM, CardioGAN, ABPNet, and PPG2ABP on the corresponding tasks, and that adding more condition modalities at test time further improves generation. The authors also claim that generated signals, without task-specific fine-tuning, restore diagnostic performance on ST-change, hypertrophy, and atrial fibrillation detection and improve heart-rate and blood-pressure estimation on unseen datasets.

Load-bearing premise

The load-bearing premise is that a single shared noise schedule for all modalities at every diffusion step suffices to represent every conditional distribution; if different signals need different noise levels for stable conditioning, the unified time step is under-specified and masking cannot recover the lost per-modality noise information.

Editorial extensions

If this is right

  • A single pre-trained UniCardio model can be deployed for denoising, imputation, and translation across PPG, ECG, and BP without retraining per task.
  • Adding more condition modalities at test time consistently lowers reconstruction error, so wearable devices that record only PPG can still obtain ECG and BP waveforms.
  • Generated signals can be fed directly into downstream classifiers and estimators, restoring accuracy, sensitivity, and specificity to near ground-truth levels on unseen datasets.
  • The continual learning schedule lets the model absorb new modalities and tasks without catastrophic forgetting, so the framework can be extended beyond the initial three signals.
  • With 6 DDIM sampling steps the per-segment inference time is under 0.4 seconds, within real-time monitoring constraints.

Reading between the lines

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

  • If the unified-time-step assumption holds, the same architecture should generalize to other correlated physiological signal families, such as respiratory effort, oxygen saturation, and EEG, without redesigning the conditioning mechanism.
  • The observed ECG-to-BP translation beating PPG-to-BP translation on both datasets suggests that multi-task pretraining may unlock latent predictive features in ECG that conventional estimators ignore; this is a testable hypothesis about representation learning rather than a physiological claim.
  • A natural stress test is whether the model remains robust on ambulatory wearable data with motion artifacts and sensor displacement, since the current benchmarks come from controlled clinical recordings; the architecture already includes entropy-based filtering, but domain shift may require adaptation.
  • If per-modality noise levels matter, the unified time step may under-specify the model; a decisive comparison would train a variant with separate time steps and compare fidelity on the same translation tasks.
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 manuscript proposes UniCardio, a unified diffusion-transformer framework for cardiovascular signal generation over PPG, ECG, and BP. It combines modality-specific encoders/decoders, task-specific attention masks, an auxiliary modality, and a continual-learning schedule to handle denoising, imputation, and cross-modal translation tasks with varying numbers of condition modalities. The model is pretrained on the Cuffless BP dataset and evaluated on generation quality (RMSE, MAE, KS-Test) and on downstream tasks (ST-change, hypertrophy, atrial fibrillation detection, and heart-rate and blood-pressure estimation) over several public datasets. The authors claim that UniCardio clearly outperforms task-specific baselines and that its generated signals match ground-truth signals in downstream applications.

Significance. If the central claims are fully supported, UniCardio would be a useful unified model for cardiovascular signal restoration and translation, with a parameter-efficient design and clear practical motivation. The paper has concrete strengths: it reports extensive experiments on multiple public datasets, includes ablations of the continual-learning components, provides clinician validation of generated ECG abnormalities, studies DDIM sampling efficiency, and states that implementation code is included in the supplementary materials. These assets support reproducibility in principle. However, the headline performance claim is not supported for the pretrained model on several benchmark tasks, and the stated methodological novelty (the unified diffusion time step) is not actually reflected in the training algorithm as written. The current manuscript therefore does not yet substantiate its central contributions.

major comments (3)
  1. [Abstract and Table 1] The abstract claims that UniCardio 'clearly outperforms recent task-specific baselines' across denoising, imputation, and translation. Table 1 does not support this for the multi-modal pretrained model: for PPG imputation, UniCardio has RMSE 0.1146 versus PulseImpute 0.0763; for ECG imputation, 0.1756 versus PulseImpute 0.1391; and for PPG-to-BP translation, 10.1538 versus ABPNet 7.2994 and PPG2ABP 6.1882. Only the fine-tuned (UniCardio-F) or multi-condition (UniCardio-M) variants are consistently better than the baselines. The abstract and the 'competing or even better' statement in Section 2.2 should be qualified, or the pretrained model's performance should be argued separately from the enhanced variants.
  2. [Sec. 4.2 and Algorithm 1] The unified-time-step design is presented as the key methodological innovation, but Algorithm 1 does not implement it. The text states that modality-specific time steps are simplified from pθ(yt−1|xtx, tx, yty, ty, ztz, tz) to pθ(yt−1|xt, yt, zt, t). However, Algorithm 1 samples clean or degraded condition signals into a condition vector c, then corrupts only the target x0 to xt and minimizes ||ϵθ(xt, c, t) − ϵt||². The condition c is never noised to the same shared time step, so the implemented model is a standard conditional denoiser with condition information at its own (clean or fixed-degradation) noise level. Either the algorithm and sampling description omit a co-noising step, or the unified time step is just a label on t; in both cases the central methodological claim is under-specified and needs code-level verification or a corrected description.
  3. [Sec. 4.3 and Algorithm 1] The role of the auxiliary modality (AM) is inconsistent between the architecture description and the training procedure. For restoration tasks, Sec. 4.3 says that AM acts as the target modality in the attention mask (j=k), receives non-informative input, and reuses the corresponding modality's decoder weights. But Algorithm 1 for restoration constructs condition c from the noisy/missing target and any other modalities, corrupts the actual target x0 to xt, and computes the loss against that corrupted real modality, with no mention of an AM diffusion target or an AM input in the loss. The paper should clarify whether AM is an actual model input and target or a bookkeeping device; as written, the architecture and the training loss describe two different mechanisms, which prevents reproduction.
minor comments (5)
  1. [Sec. 2.3] The word 'applited' in the sentence 'Although UniCardio pretrained on Cuffless BP can be directly applited to MIMIC' should be 'applied'.
  2. [Fig. 4 caption] The caption states 'diastolic BP (SBP)' but should read 'diastolic BP (DBP)'.
  3. [Fig. 4 headings] The heading 'Hypetrophy' should be spelled 'Hypertrophy'.
  4. [Table 3 caption] The label 'GP Algorithm' is used for the Gerchberg-Papoulis algorithm; this abbreviation is potentially confusing because 'GP' usually denotes Gaussian process, so the method should be named in full.
  5. [Sec. 4.4] The subsection heading 'T raining Regime' contains a spacing error and should be 'Training Regime'.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: UniCardio's generation quality is benchmarked on held-out and external data, and its self-citations are background, not load-bearing.

full rationale

UniCardio is an empirical supervised generative-modeling paper. Its central predictions—denoising/imputation/translation RMSE/MAE/KS-Test and downstream detection/estimation metrics—are evaluated on held-out test segments of the Cuffless BP dataset and on separate public datasets (PTBXL, MIMIC PERform AF, WESAD, MIMIC). Most downstream evaluations are explicitly tuning-free, and the SBP/DBP fine-tuning is stated as necessary and assessed on a strictly held-out MIMIC test set with no subject overlap. No parameter is fitted to the benchmark quantities and then relabeled as a prediction; the model inputs (PPG/ECG/BP recordings) are not constructed from the reported outputs. The 'unified time step' claim in Sec. 4.2 is under-specified relative to Algorithm 1, which corrupts only the target and passes conditions at their own degradation levels, but this is a reproducibility/specification concern rather than a circular reduction, because the reported numbers do not follow by construction from the objective. Self-citations (refs. 22, 24, 25, 28, 30, 32) serve as background for diffusion transformers and continual learning and do not justify the measured outcomes; ablations (Fig. 2c-e, Supp. Table 1) and comparisons with external baselines (Table 1) provide independent empirical support. I found no step in which an equation is definitionally equivalent to its input or a fitted parameter is renamed as a prediction.

Assumptions & free parameters 5 free parameters · 5 assumptions · 1 invented entities

The central model depends on a set of hand-chosen thresholds, schedules, and an architectural placeholder. None of these are fitted to a held-out objective in a way that invalidates the results, but each is a degree of freedom that could interact with the reported gains, and the paper provides no sensitivity analysis.

free parameters (5)
  • Entropy filtering thresholds = PPG/BP 0.2, ECG 0.3
    Used in Sec. 4.4 to remove segments before training and evaluation; values were read from the training distribution and qualitative inspection, so they shape the data the entire study depends on.
  • Phase batch allocation ratios = Phase 1: 100% 1-con; Phase 2: 50/50 1-con/2-con; Phase 3: 25/25/50; Phase 4: equal
    Chosen by hand in Sec. 4.4 to balance task exposure; no sensitivity analysis across different schedules is provided.
  • Learning rate schedule = 1e-3, 1e-4, 1e-5 with decay at 0.7*e/k and final phase
    Empirically chosen to stabilize continual learning; the paper reports e=800 'already achieves superior performance' without a search.
  • Total epochs = 800
    Selected empirically; no ablation over e is shown.
  • Denoising SNR = 15 dB
    Used to synthesize noisy inputs for denoising tasks; the choice of 15 dB is not varied, so robustness to different noise levels is untested.
assumptions (5)
  • standard math The DDPM/DDIM forward and reverse processes provide a valid approximation of the data distribution for 1D physiological signals.
    Invoked in Sec. 4.2 and Supplementary Sec. A; standard diffusion theory used without proof.
  • domain assumption PPG, ECG, and BP share enough cardiovascular information that all conditional distributions can be learned from 339 hours of ICU recordings.
    The whole multi-task premise assumes sufficient correlation and data volume; Sec. 2.1 and Sec. 4.1 rest on this.
  • ad hoc to paper A single unified diffusion time step across all modalities is sufficient to capture inter-modal condition-target relations.
    Introduced in Sec. 4.2 to simplify modality-specific time steps; if false, the conditioning would be under-specified.
  • ad hoc to paper Task-specific attention masks fully determine the task and prevent interference between tasks.
    The masking design in Eq. (5) and Sec. 4.3 is an architectural choice assumed sufficient for task isolation; the ablation shows removing masks destroys performance, but no analysis shows masks are optimal.
  • ad hoc to paper The auxiliary modality (AM) can carry the target signal's diffusion trajectory while receiving non-informative input.
    Introduced in Sec. 4.3; this placeholder is a model construct and its adequacy is assumed.
invented entities (1)
  • Auxiliary modality (AM)
    purpose: A fourth input channel that acts as the target placeholder in restoration tasks, letting the same conditioned diffusion process be used for denoising and imputation.
    AM has no physiological counterpart; it is a modeling device. There is no independent falsifiable handle, and its necessity is only demonstrated internally via ablation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Versatile Cardiovascular Signal Generation with a Unified Diffusion Transformer." pith.science (2026). https://pith.science/paper/YMISX4OE

@misc{pith2026250522306,
  author       = {Pith},
  title        = {Pith review of: Versatile Cardiovascular Signal Generation with a Unified Diffusion Transformer},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YMISX4OE}},
  note         = {Machine review of arXiv:2505.22306}
}
read the original abstract

Cardiovascular signals such as photoplethysmography (PPG), electrocardiography (ECG), and blood pressure (BP) are inherently correlated and complementary, together reflecting the health of cardiovascular system. However, their joint utilization in real-time monitoring is severely limited by diverse acquisition challenges from noisy wearable recordings to burdened invasive procedures. Here we propose UniCardio, a multi-modal diffusion transformer that reconstructs low-quality signals and synthesizes unrecorded signals in a unified generative framework. Its key innovations include a specialized model architecture to manage the signal modalities involved in generation tasks and a continual learning paradigm to incorporate varying modality combinations. By exploiting the complementary nature of cardiovascular signals, UniCardio clearly outperforms recent task-specific baselines in signal denoising, imputation, and translation. The generated signals match the performance of ground-truth signals in detecting abnormal health conditions and estimating vital signs, even in unseen domains, while ensuring interpretability for human experts. These advantages position UniCardio as a promising avenue for advancing AI-assisted healthcare.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

96 extracted references · 72 canonical work pages

  1. [1]

    https://www.who

    Organization, W.H.: Cardiovascular Diseases (CVDs). https://www.who. int/health-topics/cardiovascular-diseases (2021)

    2021

  2. [2]

    Best Practice & Research Clinical Anaesthesiology 28(4), 395–406 (2014)

    Alian, A.A., Shelley, K.H.: Photoplethysmography. Best Practice & Research Clinical Anaesthesiology 28(4), 395–406 (2014)

    2014

  3. [3]

    Heart Disease 1, 82–128 (2001)

    Mirvis, D.M., Goldberger, A.L.: Electrocardiography. Heart Disease 1, 82–128 (2001)

    2001

  4. [4]

    Critical Care Nurse 22(2), 60–79 (2002)

    McGhee, B.H., Bridges, E.J.: Monitoring arterial blood pressure: what you may not know. Critical Care Nurse 22(2), 60–79 (2002)

    2002

  5. [5]

    Communications Medicine 4(1), 1–7 (2024)

    Elgendi, M., Haugg, F., Fletcher, R.R., Allen, J., Shin, H., Alian, A., Menon, C.: Recommendations for evaluating photoplethysmography- based algorithms for blood pressure assessment. Communications Medicine 4(1), 1–7 (2024)

    2024

  6. [6]

    Electronics 3(2), 282–302 (2014)

    Tamura, T., Maeda, Y., Sekine, M., Yoshida, M.: Wearable photo- plethysmographic sensors—past and present. Electronics 3(2), 282–302 (2014)

    2014

  7. [7]

    Circulation 115(10), 1306–1324 (2007)

    Kligfield, P., Gettes, L.S., Bailey, J.J., Childers, R., Deal, B.J., Han- cock, E.W., Van Herpen, G., Kors, J.A., Macfarlane, P., Mirvis, D.M., et al.: Recommendations for the standardization and interpretation of the electrocardiogram. Circulation 115(10), 1306–1324 (2007)

    2007

  8. [8]

    Trobec, R., Tomaˇ si´ c, I., Rashkovska, A., Depolli, M., Avbelj, V.: Body sensors and electrocardiography (2018)

    2018

Show all 96 references
  1. [9]

    In: Le Roux, P.D., Levine, J.M., Kofke, W.A

    Pascual, J.L., Horak, J., Gracias, V.H., Neligan, P.J.: Chapter 19 - volume status and cardiac function. In: Le Roux, P.D., Levine, J.M., Kofke, W.A. (eds.) Monitoring in Neurocritical Care, pp. 176–1883. W.B. Saunders, Philadelphia (2013) 24

    2013

  2. [10]

    Best Practice & Research Clinical Anaesthesiology 28(4), 309–322 (2014)

    Saugel, B., Dueck, R., Wagner, J.Y.: Measurement of blood pressure. Best Practice & Research Clinical Anaesthesiology 28(4), 309–322 (2014)

    2014

  3. [11]

    IEEE Access 7, 60806–60813 (2019)

    Chiang, H.-T., Hsieh, Y.-Y., Fu, S.-W., Hung, K.-H., Tsao, Y., Chien, S.-Y.: Noise reduction in ecg signals using fully convolutional denoising autoencoders. IEEE Access 7, 60806–60813 (2019)

    2019

  4. [12]

    IEEE Sensors Letters 7(7), 1–4 (2023)

    Ahmed, R., Mehmood, A., Rahman, M.M.U., Dobre, O.A.: A deep learn- ing and fast wavelet transform-based hybrid approach for denoising of ppg signals. IEEE Sensors Letters 7(7), 1–4 (2023)

    2023

  5. [13]

    Advances in Neural Information Processing Systems 35, 26874–26888 (2022)

    Xu, M., Moreno, A., Nagesh, S., Aydemir, V., Wetter, D., Kumar, S., Rehg, J.M.: Pulseimpute: A novel benchmark task for pulsative physi- ological signal imputation. Advances in Neural Information Processing Systems 35, 26874–26888 (2022)

    2022

  6. [14]

    Proceedings of the VLDB Endowment 14(11), 2533–2545 (2021)

    Bansal, P., Deshpande, P., Sarawagi, S.: Missing value imputation on mul- tidimensional time series. Proceedings of the VLDB Endowment 14(11), 2533–2545 (2021)

    2021

  7. [15]

    In: Proceedings of the AAAI Conference on Artificial Intelligence, vol

    Shome, D., Sarkar, P., Etemad, A.: Region-disentangled diffusion model for high-fidelity ppg-to-ecg translation. In: Proceedings of the AAAI Conference on Artificial Intelligence, vol. 38, pp. 15009–15019 (2024)

    2024

  8. [16]

    In: Pro- ceedings of the AAAI Conference on Artificial Intelligence, vol

    Sarkar, P., Etemad, A.: Cardiogan: Attentive generative adversarial network with dual discriminators for synthesis of ecg from ppg. In: Pro- ceedings of the AAAI Conference on Artificial Intelligence, vol. 35, pp. 488–496 (2021)

    2021

  9. [17]

    Communications of the ACM 63(11), 139–144 (2020)

    Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., Courville, A., Bengio, Y.: Generative adversarial networks. Communications of the ACM 63(11), 139–144 (2020)

    2020

  10. [18]

    Advances in Neural Information Processing Systems 34, 17695–17709 (2021)

    De Bortoli, V., Thornton, J., Heng, J., Doucet, A.: Diffusion schr¨ odinger bridge with applications to score-based generative modeling. Advances in Neural Information Processing Systems 34, 17695–17709 (2021)

    2021

  11. [19]

    IEEE Transactions on Biomedical Engineering 64(4), 859–869 (2016)

    Kachuee, M., Kiani, M.M., Mohammadzade, H., Shabany, M.: Cuffless blood pressure estimation algorithms for continuous health-care mon- itoring. IEEE Transactions on Biomedical Engineering 64(4), 859–869 (2016)

    2016

  12. [20]

    In: Advances in Neural Information Processing Systems (2020)

    Ho, J., Jain, A., Abbeel, P.: Denoising diffusion probabilistic models. In: Advances in Neural Information Processing Systems (2020)

    2020

  13. [21]

    In: Proceedings of the International Conference on Learning Representations (2021)

    Song, Y., Sohl-Dickstein, J., Kingma, D.P., Kumar, A., Ermon, S., Poole, B.: Score-based generative modeling through stochastic differential 25 equations. In: Proceedings of the International Conference on Learning Representations (2021)

    2021

  14. [22]

    In: Proceedings of the International Conference on Machine Learning, pp

    Bao, F., Nie, S., Xue, K., Li, C., Pu, S., Wang, Y., Yue, G., Cao, Y., Su, H., Zhu, J.: One transformer fits all distributions in multi-modal diffu- sion at scale. In: Proceedings of the International Conference on Machine Learning, pp. 1692–1717 (2023). PMLR

    2023

  15. [23]

    In: Pro- ceedings of the IEEE/CVF International Conference on Computer Vision, pp

    Peebles, W., Xie, S.: Scalable diffusion models with transformers. In: Pro- ceedings of the IEEE/CVF International Conference on Computer Vision, pp. 4195–4205 (2023)

    2023

  16. [24]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (2023)

    Bao, F., Nie, S., Xue, K., Cao, Y., Li, C., Su, H., Zhu, J.: All are worth words: A vit backbone for diffusion models. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (2023)

    2023

  17. [25]

    IEEE Transactions on Pattern Analysis and Machine Intelligence (2024)

    Wang, L., Zhang, X., Su, H., Zhu, J.: A comprehensive survey of continual learning: theory, method and application. IEEE Transactions on Pattern Analysis and Machine Intelligence (2024)

    2024

  18. [26]

    Neural Networks 113, 54–71 (2019)

    Parisi, G.I., Kemker, R., Part, J.L., Kanan, C., Wermter, S.: Continual lifelong learning with neural networks: A review. Neural Networks 113, 54–71 (2019)

    2019

  19. [27]

    Proceedings of the National Academy of Sciences 114(13), 3521–3526 (2017)

    Kirkpatrick, J., Pascanu, R., Rabinowitz, N., Veness, J., Desjardins, G., Rusu, A.A., Milan, K., Quan, J., Ramalho, T., Grabska-Barwinska, A., et al.: Overcoming catastrophic forgetting in neural networks. Proceedings of the National Academy of Sciences 114(13), 3521–3526 (2017)

    2017

  20. [28]

    Advances in Neural Information Processing Systems 34, 22379–22391 (2021)

    Wang, L., Zhang, M., Jia, Z., Li, Q., Bao, C., Ma, K., Zhu, J., Zhong, Y.: Afec: Active forgetting of negative transfer in continual learning. Advances in Neural Information Processing Systems 34, 22379–22391 (2021)

    2021

  21. [29]

    In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp

    Rebuffi, S.-A., Kolesnikov, A., Sperl, G., Lampert, C.H.: Icarl: Incremen- tal classifier and representation learning. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 2001–2010 (2017)

    2017

  22. [30]

    In: Proceedings of the International Conference on Learning Representations (2021)

    Wang, L., Zhang, X., Yang, K., Yu, L., Li, C., Hong, L., Zhang, S., Li, Z., Zhong, Y., Zhu, J.: Memory replay with data compression for contin- ual learning. In: Proceedings of the International Conference on Learning Representations (2021)

    2021

  23. [31]

    In: Proceedings of the International Conference on Machine Learning, pp

    Serra, J., Suris, D., Miron, M., Karatzoglou, A.: Overcoming catas- trophic forgetting with hard attention to the task. In: Proceedings of the International Conference on Machine Learning, pp. 4548–4557 (2018) 26

    2018

  24. [32]

    Nature Machine Intelligence 5(12), 1356–1368 (2023)

    Wang, L., Zhang, X., Li, Q., Zhang, M., Su, H., Zhu, J., Zhong, Y.: Incor- porating neuro-inspired adaptability for continual learning in artificial intelligence. Nature Machine Intelligence 5(12), 1356–1368 (2023)

    2023

  25. [33]

    In: IEEE Interna- tional Instrumentation and Measurement Technology Conference, pp

    Paviglianiti, A., Randazzo, V., Pasero, E., Vallan, A.: Noninvasive arterial blood pressure estimation using abpnet and vital-ecg. In: IEEE Interna- tional Instrumentation and Measurement Technology Conference, pp. 1–5 (2020). IEEE

    2020

  26. [34]

    Bioengineering 9(11), 692 (2022)

    Ibtehaz, N., Mahmud, S., Chowdhury, M.E., Khandakar, A., Salman Khan, M., Ayari, M.A., Tahir, A.M., Rahman, M.S.: Ppg2abp: Translating photoplethysmogram (ppg) signals to arterial blood pressure (abp) waveforms. Bioengineering 9(11), 692 (2022)

    2022

  27. [35]

    Frontiers in Cardiovascular Medicine 11, 1277123 (2024)

    Bing, P., Liu, W., Zhai, Z., Li, J., Guo, Z., Xiang, Y., He, B., Zhu, L.: A novel approach for denoising electrocardiogram signals to detect cardiovascular diseases using an efficient hybrid scheme. Frontiers in Cardiovascular Medicine 11, 1277123 (2024)

    2024

  28. [36]

    IEEE signal processing letters 9(12), 438–441 (2003)

    Sendur, L., Selesnick, I.W.: Bivariate shrinkage with local variance estimation. IEEE signal processing letters 9(12), 438–441 (2003)

    2003

  29. [37]

    Optica Acta: International Journal of Optics 21(9), 709–720 (1974)

    Gerchberg, R.: Super-resolution through error energy reduction. Optica Acta: International Journal of Optics 21(9), 709–720 (1974)

    1974

  30. [38]

    IEEE Transactions on Circuits and Systems22(9), 735–742 (2003)

    Papoulis, A.: A new algorithm in spectral analysis and band-limited extrapolation. IEEE Transactions on Circuits and Systems22(9), 735–742 (2003)

    2003

  31. [39]

    Scientific Data 7(1), 1–15 (2020)

    Wagner, P., Strodthoff, N., Bousseljot, R.-D., Kreiseler, D., Lunze, F.I., Samek, W., Schaeffter, T.: Ptb-xl, a large publicly available electrocar- diography dataset. Scientific Data 7(1), 1–15 (2020)

    2020

  32. [40]

    Physiologi- cal Measurement 43(8), 085007 (2022)

    Charlton, P.H., Kotzen, K., Mej ´ ıa-Mej ´ ıa, E., Aston, P.J., Budidha, K., Mant, J., Pettit, C., Behar, J.A., Kyriacou, P.A.: Detecting beats in the photoplethysmogram: benchmarking open-source algorithms. Physiologi- cal Measurement 43(8), 085007 (2022)

    2022

  33. [41]

    In: Proceedings of the ACM International Conference on Multimodal Interaction, pp

    Schmidt, P., Reiss, A., Duerichen, R., Marberger, C., Van Laerhoven, K.: Introducing wesad, a multimodal dataset for wearable stress and affect detection. In: Proceedings of the ACM International Conference on Multimodal Interaction, pp. 400–408 (2018)

    2018

  34. [42]

    In: Computers in Cardiology, pp

    Moody, G.B., Mark, R.G.: A database to support development and evalu- ation of intelligent intensive care monitoring. In: Computers in Cardiology, pp. 657–660 (1996). IEEE 27

    1996

  35. [43]

    Scientific Reports 13(1), 986 (2023)

    Finnegan, E., Davidson, S., Harford, M., Watkinson, P., Tarassenko, L., Villarroel, M.: Features from the photoplethysmogram and the electro- cardiogram for estimating changes in blood pressure. Scientific Reports 13(1), 986 (2023)

    2023

  36. [44]

    Journal of Electrocardiology 81, 153–155 (2023)

    Landry, C., Mukkamala, R.: Current evidence suggests that estimating blood pressure from convenient ecg waveforms alone is not viable. Journal of Electrocardiology 81, 153–155 (2023)

    2023

  37. [45]

    : 2016 esc guidelines for the management of atrial fibrillation developed in col- laboration with eacts

    Kirchhof, P., Benussi, S., Kotecha, D., Ahlsson, A., Atar, D., Casadei, B., Castella, M., Diener, H.-C., Heidbuchel, H., Hendriks, J., et al. : 2016 esc guidelines for the management of atrial fibrillation developed in col- laboration with eacts. Polish Heart Journal (Kardiolo...

    2016

  38. [46]

    Circulation 130(23), 2071–2104 (2014)

    January, C.T., Wann, L.S., Alpert, J.S., Calkins, H., Cigarroa, J.E., Cleve- land Jr, J.C., Conti, J.B., Ellinor, P.T., Ezekowitz, M.D., Field, M.E., et al.: 2014 aha/acc/hrs guideline for the management of patients with atrial fibrillation: executive summary: a report of the ...

    2014

  39. [47]

    Elsevier Health Sciences, ??? (2017)

    Zipes, D.P., Jalife, J., Stevenson, W.G.: Cardiac Electrophysiology: from Cell to Bedside E-book. Elsevier Health Sciences, ??? (2017)

    2017

  40. [48]

    Circulation 138(20), 618–651 (2018)

    Thygesen, K., Alpert, J.S., Jaffe, A.S., Chaitman, B.R., Bax, J.J., Morrow, D.A., White, H.D., on behalf of the Joint European Society of Cardiology (ESC)/American College of Cardiology (ACC)/American Heart Association (AHA)/World Heart Federation (WHF) Task Force for the Univ...

    2018

  41. [49]

    Jama 284(7), 835–842 (2000)

    Antman, E.M., Cohen, M., Bernink, P.J., McCabe, C.H., Horacek, T., Papuchis, G., Mautner, B., Corbalan, R., Radley, D., Braunwald, E.: The timi risk score for unstable angina/non–st elevation mi: a method for prognostication and therapeutic decision making. Jama 284(7), 835–842 (2000)

    2000

  42. [50]

    Journal of the American College of Cardiology 64(24), 139–228 (2014) 28

    Amsterdam, E.A., Wenger, N.K., Brindis, R.G., Casey, D.E., Ganiats, T.G., Holmes, D.R., Jaffe, A.S., Jneid, H., Kelly, R.F., Kontos, M.C., et al.: 2014 aha/acc guideline for the management of patients with non–st- elevation acute coronary syndromes: a report of the american co...

    2014

  43. [51]

    In: IEEE International Colloquium on Signal Processing and Its Applications, pp

    Karthikeyan, P., Murugappan, M., Yaacob, S.: A review on stress induce- ment stimuli for assessing human stress using physiological signals. In: IEEE International Colloquium on Signal Processing and Its Applications, pp. 420–425 (2011). IEEE

    2011

  44. [52]

    IEEE Transactions on Industrial Electronics 58(10), 4857–4865 (2011)

    de Santos Sierra, A., ´Avila, C.S., Casanova, J.G., Del Pozo, G.B.: A stress- detection system based on physiological signals and fuzzy logic. IEEE Transactions on Industrial Electronics 58(10), 4857–4865 (2011)

    2011

  45. [53]

    In: Annual International Conference of the IEEE Engineering in Medicine and Biology Society, pp

    Momeni, N., Dell’Agnola, F., Arza, A., Atienza, D.: Real-time cognitive workload monitoring based on machine learning using physiological sig- nals in rescue missions. In: Annual International Conference of the IEEE Engineering in Medicine and Biology Society, pp. 3779–3785 (2...

    2019

  46. [54]

    Neuroscience & Biobehavioral Reviews 80, 516–537 (2017)

    Ranchet, M., Morgan, J.C., Akinwuntan, A.E., Devos, H.: Cognitive work- load across the spectrum of cognitive impairments: A systematic review of physiological measures. Neuroscience & Biobehavioral Reviews 80, 516–537 (2017)

    2017

  47. [55]

    Sensors 18(7), 2074 (2018)

    Shu, L., Xie, J., Yang, M., Li, Z., Li, Z., Liao, D., Xu, X., Yang, X.: A review of emotion recognition using physiological signals. Sensors 18(7), 2074 (2018)

    2018

  48. [56]

    Journal of Medical and Biological Engineering 40, 149–157 (2020)

    Ayata, D., Yaslan, Y., Kamasak, M.E.: Emotion recognition from multi- modal physiological signals for emotion aware healthcare systems. Journal of Medical and Biological Engineering 40, 149–157 (2020)

    2020

  49. [57]

    Cognitive Neurodynamics 14(5), 709–721 (2020)

    Xuan, Q., Wu, J., Shen, J., Ji, X., Lyu, Y., Zhang, Y.: Assessing cog- nitive load in adolescent and adult students using photoplethysmogram morphometrics. Cognitive Neurodynamics 14(5), 709–721 (2020)

    2020

  50. [58]

    Scientific Reports 13(1), 22592 (2023)

    Pavlov, Y.G., Gashkova, A.S., Kasanov, D., Kosachenko, A.I., Kotyusov, A.I., Kotchoubey, B.: Task-evoked pulse wave amplitude tracks cognitive load. Scientific Reports 13(1), 22592 (2023)

    2023

  51. [59]

    Medical & Biological Engineering & Computing 56(12), 2273–2286 (2018)

    Al Abdi, R.M., Alhitary, A.E., Abdul Hay, E.W., Al-Bashir, A.K.: Objec- tive detection of chronic stress using physiological parameters. Medical & Biological Engineering & Computing 56(12), 2273–2286 (2018)

    2018

  52. [60]

    Biological Psychology 121, 62–73 (2016)

    Mandrick, K., Peysakhovich, V., R´ emy, F., Lepron, E., Causse, M.: Neural and psychophysiological correlates of human performance under stress and high mental workload. Biological Psychology 121, 62–73 (2016)

    2016

  53. [61]

    Psychophysiology 45(5), 869–875 (2008)

    Richter, M., Friedrich, A., Gendolla, G.H.: Task difficulty effects on cardiac activity. Psychophysiology 45(5), 869–875 (2008)

    2008

  54. [62]

    Journal of Physiology 594(24), 7465–7482 (2016)

    El Sayed, K., Macefield, V.G., Hissen, S.L., Joyner, M.J., Taylor, C.E.: 29 Rate of rise in diastolic blood pressure influences vascular sympathetic response to mental stress. Journal of Physiology 594(24), 7465–7482 (2016)

    2016

  55. [63]

    arXiv preprint arXiv:2303.04226 (2023)

    Cao, Y., Li, S., Liu, Y., Yan, Z., Dai, Y., Yu, P.S., Sun, L.: A comprehen- sive survey of ai-generated content (aigc): A history of generative ai from gan to chatgpt. arXiv preprint arXiv:2303.04226 (2023)

    2023 arXiv

  56. [64]

    IEEE Transactions on Artificial Intelligence (2024)

    Mai, W., Zhang, J., Fang, P., Zhang, Z.: Brain-conditional multimodal synthesis: A survey and taxonomy. IEEE Transactions on Artificial Intelligence (2024)

    2024

  57. [65]

    Bioengineering 3(4), 21 (2016)

    Elgendi, M.: Optimal signal quality index for photoplethysmogram sig- nals. Bioengineering 3(4), 21 (2016)

    2016

  58. [66]

    Scientific Reports 9(1), 6734 (2019)

    Zhu, F., Ye, F., Fu, Y., Liu, Q., Shen, B.: Electrocardiogram generation with a bidirectional lstm-cnn generative adversarial network. Scientific Reports 9(1), 6734 (2019)

    2019

  59. [67]

    IEEE Journal of Biomedical and Health Informatics 26(5), 2136–2146 (2022)

    Mazumder, O., Banerjee, R., Roy, D., Bhattacharya, S., Ghose, A., Sinha, A.: Synthetic ppg signal generation to improve coronary artery disease classification: Study with physical model of cardiovascular system. IEEE Journal of Biomedical and Health Informatics 26(5), 2136–2146 (2022)

    2022

  60. [68]

    EURASIP Journal on Advances in Signal Processing 2024(1), 95 (2024)

    Ezzat, A., Omer, O.A., Mohamed, U.S., Mubarak, A.S.: Ecg signal recon- struction from ppg using a hybrid attention-based deep learning network. EURASIP Journal on Advances in Signal Processing 2024(1), 95 (2024)

    2024

  61. [69]

    Advances in Neural Information Processing Systems 29 (2016)

    Van den Oord, A., Kalchbrenner, N., Espeholt, L., Vinyals, O., Graves, A., et al.: Conditional image generation with pixelcnn decoders. Advances in Neural Information Processing Systems 29 (2016)

    2016

  62. [70]

    arXiv preprint arXiv:1609.03499 12 (2016)

    Van Den Oord, A., Dieleman, S., Zen, H., Simonyan, K., Vinyals, O., Graves, A., Kalchbrenner, N., Senior, A., Kavukcuoglu, K., et al.: Wavenet: A generative model for raw audio. arXiv preprint arXiv:1609.03499 12 (2016)

    2016 arXiv

  63. [71]

    In: Proceedings of the International Conference on Learning Representations (2021)

    Kong, Z., Ping, W., Huang, J., Zhao, K., Catanzaro, B.: Diffwave: A versatile diffusion model for audio synthesis. In: Proceedings of the International Conference on Learning Representations (2021)

    2021

  64. [72]

    Advances in Neural Information Processing Systems 34, 24804–24816 (2021)

    Tashiro, Y., Song, J., Song, Y., Ermon, S.: Csdi: Conditional score-based diffusion models for probabilistic time series imputation. Advances in Neural Information Processing Systems 34, 24804–24816 (2021)

    2021

  65. [73]

    Behavior Research Methods, 1–8 (2021)

    Makowski, D., Pham, T., Lau, Z.J., Brammer, J.C., Lespinasse, F., Pham, H., Sch¨ olzel, C., Chen, S.A.: Neurokit2: A python toolbox for 30 neurophysiological signal processing. Behavior Research Methods, 1–8 (2021)

    2021

  66. [74]

    American Journal of Physiology-Heart and Circulatory Physiology 278(6), 2039–2049 (2000)

    Richman, J.S., Moorman, J.R.: Physiological time-series analysis using approximate entropy and sample entropy. American Journal of Physiology-Heart and Circulatory Physiology 278(6), 2039–2049 (2000)

    2000

  67. [75]

    Journal of the American Heart Association 11(7), 023555 (2022)

    Liu, Z., Zhou, B., Jiang, Z., Chen, X., Li, Y., Tang, M., Miao, F.: Multiclass arrhythmia detection and classification from photoplethysmog- raphy signals using a deep convolutional neural network. Journal of the American Heart Association 11(7), 023555 (2022)

    2022

  68. [76]

    IEEE Journal of Translational Engineering in Health and Medicine 6, 1–11 (2018)

    Nabian, M., Yin, Y., Wormwood, J., Quigley, K.S., Barrett, L.F., Ostad- abbas, S.: An open-source feature extraction tool for the analysis of peripheral physiological data. IEEE Journal of Translational Engineering in Health and Medicine 6, 1–11 (2018)

    2018

  69. [77]

    Scientific Reports 11(1), 13539 (2021)

    Jeong, D.U., Lim, K.M.: Combined deep cnn–lstm network-based multi- tasking learning architecture for noninvasive continuous blood pressure estimation using difference in ecg-ppg features. Scientific Reports 11(1), 13539 (2021)

    2021

  70. [78]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (2022)

    Rombach, R., Blattmann, A., Lorenz, D., Esser, P., Ommer, B.: High- resolution image synthesis with latent diffusion models. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (2022)

    2022

  71. [79]

    In: Proceedings of the International Conference on Machine Learning (2024)

    Esser, P., Kulal, S., Blattmann, A., Entezari, R., M¨ uller, J., Saini, H., Levi, Y., Lorenz, D., Sauer, A., Boesel, F., Podell, D., Dockhorn, T., English, Z., Rombach, R.: Scaling rectified flow transformers for high-resolution image synthesis. In: Proceedings of the Internat...

    2024

  72. [80]

    In: IEEE International Conference on Acoustics, Speech and Signal Processing (2023)

    Chen, Z., Tan, X., Wang, K., Pan, S., Mandic, D.P., He, L., Zhao, S.: Infergrad: Improving diffusion models for vocoder by considering inference in training. In: IEEE International Conference on Acoustics, Speech and Signal Processing (2023)

    2023

  73. [81]

    In: Proceedings of the International Conference on Machine Learning (2023)

    Liu, H., Chen, Z., Yuan, Y., Mei, X., Liu, X., Mandic, D.P., Wang, W., Plumbley, M.D.: Audioldm: Text-to-audio generation with latent diffu- sion models. In: Proceedings of the International Conference on Machine Learning (2023)

    2023

  74. [82]

    In: IEEE International Conference on Acoustics, Speech and Signal Processing (2025) 31

    Mo, S., Chen, Z., Bao, F., Zhu, J.: Diffgap: A lightweight diffusion module in contrastive space for bridging cross-model gap. In: IEEE International Conference on Acoustics, Speech and Signal Processing (2025) 31

    2025

  75. [83]

    In: Proceedings of the ACM International Conference on Multimedia (2024)

    Wang, X., Wang, Y., Wu, Y., Song, R., Tan, X., Chen, Z., Xu, H., Sui, G.: Tiva: Time-aligned video-to-audio generation. In: Proceedings of the ACM International Conference on Multimedia (2024)

    2024

  76. [84]

    In: Advances in Neural Information Processing Systems (2023)

    Wang, Z., Lu, C., Wang, Y., Bao, F., Li, C., Su, H., Zhu, J.: Prolific- dreamer: High-fidelity and diverse text-to-3d generation with variational score distillation. In: Advances in Neural Information Processing Systems (2023)

    2023

  77. [85]

    arXiv preprint arXiv:2405.04233 (2024)

    Bao, F., Xiang, C., Yue, G., He, G., Zhu, H., Zheng, K., Zhao, M., Liu, S., Wang, Y., Zhu, J.: Vidu: a highly consistent, dynamic and skilled text-to- video generator with diffusion models. arXiv preprint arXiv:2405.04233 (2024)

    2024 arXiv

  78. [86]

    In: Proceedings of the International Conference on Machine Learning (2025)

    Wang, Y., Chen, Z., Xiaoyu, C., Wei, Y., Zhu, J., Chen, J.: Framebridge: Improving image-to-video generation with bridge models. In: Proceedings of the International Conference on Machine Learning (2025)

    2025

  79. [87]

    In: IEEE International Conference on Acoustics, Speech and Signal Processing (2025)

    Miao, Y., Chen, Z., Li, C., Mandic, D.P.: Respdiff: An end-to-end multi- scale RNN diffusion model for respiratory waveform estimation from PPG signals. In: IEEE International Conference on Acoustics, Speech and Signal Processing (2025)

    2025

  80. [88]

    arXiv preprint arXiv:2310.15742 (2023)

    Jenkins, A., Chen, Z., Ng, F.S., Mandic, D.P.: Improving diffusion models for ECG imputation with an augmented template prior. arXiv preprint arXiv:2310.15742 (2023)

    2023 arXiv

  81. [89]

    Chang, P., Li, H., Quan, S.F., Lu, S., Wung, S., Roveda, J., Li, A.: A transformer-based diffusion probabilistic model for heart rate and blood pressure forecasting in intensive care unit. Comput. Methods Programs Biomed. 246, 108060 (2024)

    2024

  82. [90]

    In: Proceedings of the International Conference on Learning Representations (2021)

    Song, J., Meng, C., Ermon, S.: Denoising diffusion implicit models. In: Proceedings of the International Conference on Learning Representations (2021)

    2021

  83. [91]

    IEEE Internet Things J

    Ma, C., Guo, L., Zhang, H., Liu, Z., Zhang, G.: Diffcnbp: Lightweight dif- fusion model for iomt-based continuous cuffless blood pressure waveform monitoring using PPG. IEEE Internet Things J. 12(1), 61–80 (2025)

    2025

  84. [92]

    In: Proceedings of the International Conference on Learning Representations (2024)

    Zhou, L., Lou, A., Khanna, S., Ermon, S.: Denoising diffusion bridge models. In: Proceedings of the International Conference on Learning Representations (2024)

    2024

  85. [93]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision (2023) 32

    Hang, T., Gu, S., Li, C., Bao, J., Chen, D., Hu, H., Geng, X., Guo, B.: Efficient diffusion training via min-snr weighting strategy. In: Proceedings of the IEEE/CVF International Conference on Computer Vision (2023) 32

    2023

  86. [94]

    In: Proceedings of the International Conference on Machine Learning (2021)

    Popov, V., Vovk, I., Gogoryan, V., Sadekova, T., Kudinov, M.A.: Grad- tts: A diffusion probabilistic model for text-to-speech. In: Proceedings of the International Conference on Machine Learning (2021)

    2021

  87. [95]

    Bao, F., Li, C., Sun, J., Zhu, J.: Why are conditional generative models better than unconditional ones? arXiv preprint arXiv:2212.00362 (2022)

    2022 arXiv

  88. [96]

    -N” denotes the noised signals. “-I

    Chen, Z., Dees, B.S., Mandic, D.P.: A probabilistic beat-to-beat filter- ing model for continuous and accurate blood pressure estimation. In: Proceedings of the International Joint Conference on Neural Networks (2020) 33 A Diffusion Models A.1 Unconditional Diffusion Models Di...

    2020

Pith tools

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