Results explorer

Name: ChatPPG wearable AI results
Creator: ChatPPG
License: https://creativecommons.org/licenses/by/4.0/

One row per reported result from wearable PPG, ECG, motion and sleep papers. Filter by task, dataset, metric, sensor and year, or show only SOTA rows. Each value links to its source: a paper table or figure, or the leaderboard row it came from.

Last updated Oct 27, 2026

433
reported results
224
with a table or figure reference
92
papers and models
134
task × dataset × metric pairs
38
benchmarks with a SOTA star

SOTA only

433 results

Reported results with task, dataset, metric, value and source
			Metric		Comparability	Sensor / site		Source	Links
Chronos Baseline in PaPaGei: Open Foundation Models for Optical Physiological Signals	Apnea-hypopnea index	MESA	MAE (events/h)	14.06↓	subject-level train/val/testwithin-datasetlinear probe on frozen embeddings	PPG	2024	Tables 3 and 4 (as reported) (opens in a new tab)
SimCLR Baseline in PaPaGei: Open Foundation Models for Optical Physiological Signals	Apnea-hypopnea index	MESA	MAE (events/h)	14.17↓	subject-level train/val/testwithin-datasetlinear probe on frozen embeddings	PPG	2020	Tables 3 and 4 (as reported) (opens in a new tab)
Moment Baseline in PaPaGei: Open Foundation Models for Optical Physiological Signals	Apnea-hypopnea index	MESA	MAE (events/h)	14.23↓	subject-level train/val/testwithin-datasetlinear probe on frozen embeddings	PPG	2024	Tables 3 and 4 (as reported) (opens in a new tab)
BYOL Baseline in PaPaGei: Open Foundation Models for Optical Physiological Signals	Apnea-hypopnea index	MESA	MAE (events/h)	14.26↓	subject-level train/val/testwithin-datasetlinear probe on frozen embeddings	PPG	2020	Tables 3 and 4 (as reported) (opens in a new tab)
TF-C Baseline in PaPaGei: Open Foundation Models for Optical Physiological Signals	Apnea-hypopnea index	MESA	MAE (events/h)	15.10↓	subject-level train/val/testwithin-datasetlinear probe on frozen embeddings	PPG	2022	Tables 3 and 4 (as reported) (opens in a new tab)
Stat. features + RF Baseline in PaPaGei: Open Foundation Models for Optical Physiological Signals	Apnea-hypopnea index	MESA	MAE (events/h)	15.31↓	subject-level train/val/testwithin-datasetlinear probe on frozen embeddings	PPG	—	Tables 3 and 4 (as reported) (opens in a new tab)
REGLE Baseline in PaPaGei: Open Foundation Models for Optical Physiological Signals	Apnea-hypopnea index	MESA	MAE (events/h)	15.54↓	subject-level train/val/testwithin-datasetlinear probe on frozen embeddings	PPG	2024	Tables 3 and 4 (as reported) (opens in a new tab)
Random Forest (engineered features) In Development of an efficient deep learning model for arrhythmia detection from consumer smartwatches	Arrhythmia classification	Pulsewatch	Macro F1 (%)	84.17↑	two-fold subject-independent cross-validationwithin-datasetclassical machine learning on PPG + accelerometer featuresNSR / AF / PAC-PVC	PPG, ACC	2026	Abstract (Results) (opens in a new tab)
CogAdapt In CogAdapt: Adapting Clinical ECG Foundation Models for Wearable Cognitive Load Assessment	Cognitive load	CL-Drive	Macro F1	0.768↑	leave-one-subject-out cross-validationwithin-datasetadapted clinical ECG foundation model	ECG	2026	Table I (opens in a new tab)
CogAdapt In CogAdapt: Adapting Clinical ECG Foundation Models for Wearable Cognitive Load Assessment	Cognitive load	CLARE	Macro F1	0.626↑	leave-one-subject-out cross-validationwithin-datasetadapted clinical ECG foundation model	ECG	2026	Table I (opens in a new tab)
Awake–sleep PPG indices + SVM Hypertension screening via awake–sleep differences in photoplethysmogram signals	Cuffless blood pressure	aurora-bp-htn	Accuracy	0.811↑	Protocol not structured	—	2026	Leaderboard row	Leaderboard
XGBoost on interpretable PPG + demographic featu Cross-Dataset Evaluation of Interpretable Photoplethysmography Features for Blood Pressure Category Classification	Cuffless blood pressure	aurora-bp	Area under ROC	0.778↑	Protocol not structured	PPG	2026	Leaderboard row	Leaderboard
PC-AVCT wearable-state model Context-Augmented, Session-Calibrated Blood Pressure Tracking from Photoplethysmography	Cuffless blood pressure	aurora-bp	Mean absolute error (mmHg); cuffless BP	7.697↓	Protocol not structured	PPG, IMU · wrist	2026	Leaderboard row	Leaderboard
PulseAnchor PulseAnchor: Personalizing PPG Foundation Models for Cuffless Blood Pressure Estimation	Cuffless blood pressure	aurora-bp	Mean absolute error (mmHg); cuffless BP	7.84↓	uses extra training data	PPG · wrist	2026	Leaderboard row	Leaderboard
Hybrid features + personalized calibration In Personalized Calibration and Hybrid Feature Fusion for Continuous Blood Pressure Estimation Using PPG and ECG Signals	Cuffless blood pressure	BP-UCI	MAE (mmHg)	7.85↓	not stated in the abstractwithin-datasetfeature-based model with base-BP calibrationSBP	PPG, ECG	2026	Abstract (Results) (opens in a new tab)
Awake–sleep PPG temporal indices + SVM Hypertension screening via awake–sleep differences in photoplethysmogram signals	Cuffless blood pressure	cuhk-bp-htn	Accuracy	0.846↑	Protocol not structured	—	2026	Leaderboard row	Leaderboard
3M-BPNet Utilizing Multi-PPG-Sensor Site Information in a Localized Wrist Area for Improving Cuffless Blood Pressure Estimation	Cuffless blood pressure	liu-wrist-9ch-array-121	Mean absolute error (mmHg); cuffless BP	4.84↓	subjects overlap	wrist	2026	Leaderboard row	Leaderboard
McBP-Net Dynamic Beat-to-Beat Measurements of Blood Pressure Using Multimodal Physiological Signals and a Hybrid CNN-LSTM Model	Cuffless blood pressure	mcbp-cold-pressor-23	Mean absolute error (mmHg); cuffless BP	4.19↓	Protocol not structured	PPG, ECG, TEMP	2025	Leaderboard row	Leaderboard
FiLMamba In FiLMamba: A FiLM-Conditioned Mamba Architecture with an Energy-Efficient FPGA Accelerator for Generalizable Cuffless Blood Pressure Estimation	Cuffless blood pressure	MIMIC-BP (1,524 patients)	SD of error (mmHg)	6.45↓	not stated in the abstractwithin-datasetFiLM-conditioned Mamba, end-to-endSBP	PPG	2026	Abstract (Results) (opens in a new tab)
FiLMamba FiLMamba: A FiLM-Conditioned Mamba Architecture with an Energy-Efficient FPGA Accelerator for Generalizable Cuffless Blood Pressure Estimation	Cuffless blood pressure	MIMIC-BP	Standard deviation of error (mmHg); cuffless BP	6.45↓	Protocol not structured	PPG, ABP, ECG, RESP · finger	2026	Leaderboard row	Leaderboard
Direct PPG→BP pipeline In Blood Pressure Estimation from PPG: A Comparative Study of Direct and ECG-Mediated Deep Learning Pipelines	Cuffless blood pressure	MIMIC-III waveform	MAE (mmHg)	4.82↓	segment-level train/validation/test split (1.74M / 218k / 217k segments; patient separation not stated)within-datasetend-to-end training, no subject calibrationSBP	PPG	2026	Table V (opens in a new tab)
Multiresolution U-Net @62.5 Hz Practical optimization of deep learning models for cuffless blood pressure estimation from photoplethysmography	Cuffless blood pressure	mimic-waveform-205k	Mean absolute error (mmHg); cuffless BP	5.24↓	Protocol not structured	—	2026	Leaderboard row	Leaderboard
Soft, Multi-Wavelength Photoplethysmography Enab Soft, Multi-Wavelength Photoplethysmography Enables Reliable Neonatal Blood Pressure Monitoring Via Error Stratification	Cuffless blood pressure	neo-bp	Mean absolute error (mmHg); cuffless BP	9.67↓	Protocol not structured	ABP	2026	Leaderboard row	Leaderboard
McBP-Net In Dynamic Beat-to-Beat Measurements of Blood Pressure Using Multimodal Physiological Signals and a Hybrid CNN-LSTM Model	Cuffless blood pressure	Own cold-pressor recordings (23 subjects)	MAE (mmHg)	4.19↓	not stated in the abstractwithin-datasetmultimodal CNN-LSTM, beat-to-beatSBP	PPG, ECG	2025	Abstract (Results) (opens in a new tab)
Windkessel-regularized model In Cuffless Blood Pressure Estimation From Wearable Photoplethysmography and Windkessel Model-Regularized Learning Approach	Cuffless blood pressure	Own finger-wearable recordings (10 healthy subjects)	MAE (mmHg)	4.10↓	not stated in the abstractwithin-datasetphysics-regularized deep modelSBP	PPG	2026	Abstract (Results) (opens in a new tab)
3M-BPNet In Utilizing Multi-PPG-Sensor Site Information in a Localized Wrist Area for Improving Cuffless Blood Pressure Estimation	Cuffless blood pressure	Own wrist 9-channel PPG array (121 subjects)	MAE (mmHg)	4.84↓	not stated in the abstractwithin-datasetmulti-site PPG modelSBP	PPG	2026	Abstract (Results) (opens in a new tab)
TEX-MEX 0.56M PPG FM TEX-MEX	Cuffless blood pressure	PPG-BP Database (Guilin / Liang 2018)	Area under ROC	0.7347↑	uses extra training data	PPG · finger	2026	Leaderboard row	Hugging Face (opens in a new tab)Leaderboard
Chowdhury GPR + ReliefF Estimating Blood Pressure from the Photoplethysmogram Signal and Demographic Features Using Machine Learning Techniques	Cuffless blood pressure	PPG-BP Database (Guilin / Liang 2018)	Mean absolute error (mmHg); cuffless BP	3.02↓	subjects overlap	PPG · finger	2020	Leaderboard row	Leaderboard
TF-MambaNet TF-MambaNet: A Temporal and Frequency Fused Bidirectional Mamba Architecture for PPG Foundation Model	Cuffless blood pressure	PPG-BP Database (Guilin / Liang 2018)	Mean absolute error (mmHg); cuffless BP	12.89↓	uses extra training data	PPG · finger	2026	Leaderboard row	Leaderboard
AnyPPG PPG-Bench: A Unified Benchmark for PPG Foundation Models	Cuffless blood pressure	PPG-BP Database (Guilin / Liang 2018)	Mean absolute error (mmHg); cuffless BP	13.32↓	subject-disjointuses extra training data	PPG · finger	2026	Leaderboard row	Code (opens in a new tab)Leaderboard
TEX-MEX 0.56M PPG FM TEX-MEX	Cuffless blood pressure	PPG-BP Database (Guilin / Liang 2018)	Mean absolute error (mmHg); cuffless BP	14.007↓	uses extra training data	PPG · finger	2026	Leaderboard row	Hugging Face (opens in a new tab)Leaderboard
Pulse-PPG In Pulse-PPG: An Open-Source Field-Trained PPG Foundation Model for Wearable Applications Across Lab and Field Settings	Cuffless blood pressure	PPG-BP	MAE (mmHg)	8.878↓	subject-wisewithin-datasetlinear probe on frozen embeddingsdiastolic BP	PPG · finger	2025	Table 6 (opens in a new tab)	Code (opens in a new tab)
PaPaGei Baseline in Pulse-PPG: An Open-Source Field-Trained PPG Foundation Model for Wearable Applications Across Lab and Field Settings	Cuffless blood pressure	PPG-BP	MAE (mmHg)	9.246↓	subject-wisewithin-datasetlinear probe on frozen embeddingsdiastolic BP	PPG · finger	2024	Table 6 (as reported) (opens in a new tab)	Code (opens in a new tab)
MOMENT Baseline in Pulse-PPG: An Open-Source Field-Trained PPG Foundation Model for Wearable Applications Across Lab and Field Settings	Cuffless blood pressure	PPG-BP	MAE (mmHg)	9.447↓	subject-wisewithin-datasetlinear probe on frozen embeddingsdiastolic BP	PPG · finger	2024	Table 6 (as reported) (opens in a new tab)
Chronos Baseline in Pulse-PPG: An Open-Source Field-Trained PPG Foundation Model for Wearable Applications Across Lab and Field Settings	Cuffless blood pressure	PPG-BP	MAE (mmHg)	9.588↓	subject-wisewithin-datasetlinear probe on frozen embeddingsdiastolic BP	PPG · finger	2024	Table 6 (as reported) (opens in a new tab)
Dilated-residual attention U-Net In Calibration-Based Cuffless Blood Pressure Estimation Using a Dilated-Residual Attention U-Net	Cuffless blood pressure	PulseDB (calibration-based)	MAE (mmHg)	2.03↓	calibration-based subset (same subjects in training and test)within-datasetend-to-end, PPG+ECG inputDBP	PPG, ECG	2026	Abstract (Results) (opens in a new tab)
XResNet1d101 In Generalizable deep learning for photoplethysmography-based blood pressure estimation — A benchmarking study	Cuffless blood pressure	PulseDB (in-distribution, no calibration)	MAE (mmHg)	13.9↓	in-distribution test without subject-specific calibrationwithin-datasetend-to-endSBP	PPG	2025	Abstract (Results) (opens in a new tab)
GPT-PPG GPT-PPG: a GPT-based foundation model for photoplethysmography signals	Cuffless blood pressure	PulseDB v2.0	Mean absolute error (mmHg); cuffless BP	8.12↓	subject-disjointuses extra training data	PPG, ECG, ABP · finger	2025	Leaderboard row	Leaderboard
Dilated-residual attention U-Net Calibration-Based Cuffless Blood Pressure Estimation Using a Dilated-Residual Attention U-Net	Cuffless blood pressure	pulsedb-calbased	Mean absolute error (mmHg); cuffless BP	3.81↓	Protocol not structured	PPG, ECG	2026	Leaderboard row	Leaderboard
PPG-FusionNet PPG-FusionNet: A Dual-Branch Neural Architecture for Cuffless Blood Pressure Estimation from Photoplethysmography	Cuffless blood pressure	pulsedb-calbased	Mean absolute error (mmHg); cuffless BP	7.46↓	Protocol not structured	PPG · finger	2026	Leaderboard row	Leaderboard
ECG-PPG JEPA 345M Bridging ECG and PPG: Latent-Space Prediction for Robust Physiological Analysis	Cuffless blood pressure	pulsedb-vital	Mean absolute error (mmHg); cuffless BP	8.23↓	uses extra training data	finger	2026	Leaderboard row	Leaderboard
Few-shot on frozen FM embeddings Data Efficient Cuffless Blood Pressure Classification from PPG using Foundation Model Embeddings and Few-Shot Learning	Cuffless blood pressure	pulsedb-vitaldb-aha6	Accuracy	0.824↑	uses extra training data	—	2026	Leaderboard row	Leaderboard
DPIKAN DPIKAN: Decomposed Physics-Informed KANs for Physiological Signal Representation Learning	Cuffless blood pressure	ring-hgcpt	Standard deviation of error (mmHg); cuffless BP	5.41↓	subject-disjoint	PPG · ring	2026	Leaderboard row	Code (opens in a new tab)Leaderboard
ExpertoRhythm In ExpertoRhythm: Morphology-Aware Learning for Waveform Reconstruction and Cuffless Blood Pressure Estimation from Single-Channel PPG	Cuffless blood pressure	UCI cuff-less BP (942 subjects)	MAE (mmHg)	2.46↓	subject-block split: fixed 25% held-out test, 5-fold CV on the rest for model selectionwithin-datasetPPG→ABP waveform reconstruction, end-to-endSBP	PPG	2026	Table 3 (opens in a new tab)	Code (opens in a new tab)
SimCLR Baseline in AnyPPG: An ECG-Guided PPG Foundation Model Trained on Over 100,000 Hours of Recordings for Holistic Health Profiling	Cuffless blood pressure	UCI-BP	MAE (mmHg)	7.14↓	as defined per dataset in the paperwithin-datasetlinear probe on frozen embeddingsdiastolic BP	PPG	2020	Table 3 (as reported) (opens in a new tab)
AnyPPG In AnyPPG: An ECG-Guided PPG Foundation Model Trained on Over 100,000 Hours of Recordings for Holistic Health Profiling	Cuffless blood pressure	UCI-BP	MAE (mmHg)	7.14↓	as defined per dataset in the paperwithin-datasetlinear probe on frozen embeddingsdiastolic BP	PPG	2025	Table 3 (opens in a new tab)	Code (opens in a new tab)
BYOL Baseline in AnyPPG: An ECG-Guided PPG Foundation Model Trained on Over 100,000 Hours of Recordings for Holistic Health Profiling	Cuffless blood pressure	UCI-BP	MAE (mmHg)	7.31↓	as defined per dataset in the paperwithin-datasetlinear probe on frozen embeddingsdiastolic BP	PPG	2020	Table 3 (as reported) (opens in a new tab)
MOMENT Baseline in AnyPPG: An ECG-Guided PPG Foundation Model Trained on Over 100,000 Hours of Recordings for Holistic Health Profiling	Cuffless blood pressure	UCI-BP	MAE (mmHg)	7.53↓	as defined per dataset in the paperwithin-datasetlinear probe on frozen embeddingsdiastolic BP	PPG	2024	Table 3 (as reported) (opens in a new tab)
PulsePPG Baseline in AnyPPG: An ECG-Guided PPG Foundation Model Trained on Over 100,000 Hours of Recordings for Holistic Health Profiling	Cuffless blood pressure	UCI-BP	MAE (mmHg)	7.57↓	as defined per dataset in the paperwithin-datasetlinear probe on frozen embeddingsdiastolic BP	PPG	2025	Table 3 (as reported) (opens in a new tab)	Code (opens in a new tab)
ExpertoRhythm ExpertoRhythm: Morphology-Aware Learning for Waveform Reconstruction and Cuffless Blood Pressure Estimation from Single-Channel PPG	Cuffless blood pressure	uci-bp	Mean absolute error (mmHg); cuffless BP	2.46↓	Protocol not structured	PPG, ABP · finger	2026	Leaderboard row	Code (opens in a new tab)Leaderboard

Chronos

Baseline in PaPaGei: Open Foundation Models for Optical Physiological Signals

14.06↓

Apnea-hypopnea index · MESA · MAE (events/h)

subject-level train/val/testwithin-datasetlinear probe on frozen embeddings

Tables 3 and 4 (as reported)

2024

SimCLR

Baseline in PaPaGei: Open Foundation Models for Optical Physiological Signals

14.17↓

Apnea-hypopnea index · MESA · MAE (events/h)

subject-level train/val/testwithin-datasetlinear probe on frozen embeddings

Tables 3 and 4 (as reported)

2020

Moment

Baseline in PaPaGei: Open Foundation Models for Optical Physiological Signals

14.23↓

Apnea-hypopnea index · MESA · MAE (events/h)

subject-level train/val/testwithin-datasetlinear probe on frozen embeddings

Tables 3 and 4 (as reported)

2024

BYOL

Baseline in PaPaGei: Open Foundation Models for Optical Physiological Signals

14.26↓

Apnea-hypopnea index · MESA · MAE (events/h)

subject-level train/val/testwithin-datasetlinear probe on frozen embeddings

Tables 3 and 4 (as reported)

2020

TF-C

Baseline in PaPaGei: Open Foundation Models for Optical Physiological Signals

15.10↓

Apnea-hypopnea index · MESA · MAE (events/h)

subject-level train/val/testwithin-datasetlinear probe on frozen embeddings

Tables 3 and 4 (as reported)

2022

Stat. features + RF

Baseline in PaPaGei: Open Foundation Models for Optical Physiological Signals

15.31↓

Apnea-hypopnea index · MESA · MAE (events/h)

subject-level train/val/testwithin-datasetlinear probe on frozen embeddings

Tables 3 and 4 (as reported)

REGLE

Baseline in PaPaGei: Open Foundation Models for Optical Physiological Signals

15.54↓

Apnea-hypopnea index · MESA · MAE (events/h)

subject-level train/val/testwithin-datasetlinear probe on frozen embeddings

Tables 3 and 4 (as reported)

2024

Random Forest (engineered features)

In Development of an efficient deep learning model for arrhythmia detection from consumer smartwatches

84.17↑

Arrhythmia classification · Pulsewatch · Macro F1 (%)

two-fold subject-independent cross-validationwithin-datasetclassical machine learning on PPG + accelerometer featuresNSR / AF / PAC-PVC

Abstract (Results)

2026

CogAdapt

In CogAdapt: Adapting Clinical ECG Foundation Models for Wearable Cognitive Load Assessment

0.768↑

Cognitive load · CL-Drive · Macro F1

leave-one-subject-out cross-validationwithin-datasetadapted clinical ECG foundation model

Table I

2026

CogAdapt

In CogAdapt: Adapting Clinical ECG Foundation Models for Wearable Cognitive Load Assessment

0.626↑

Cognitive load · CLARE · Macro F1

leave-one-subject-out cross-validationwithin-datasetadapted clinical ECG foundation model

Table I

2026

Awake–sleep PPG indices + SVM

Hypertension screening via awake–sleep differences in photoplethysmogram signals

0.811↑

Cuffless blood pressure · aurora-bp-htn · Accuracy

Protocol not structured

Leaderboard row

2026

XGBoost on interpretable PPG + demographic featu

Cross-Dataset Evaluation of Interpretable Photoplethysmography Features for Blood Pressure Category Classification

0.778↑

Cuffless blood pressure · aurora-bp · Area under ROC

Protocol not structured

Leaderboard row

2026

PC-AVCT wearable-state model

Context-Augmented, Session-Calibrated Blood Pressure Tracking from Photoplethysmography

7.697↓

Cuffless blood pressure · aurora-bp · Mean absolute error (mmHg); cuffless BP

Protocol not structured

Leaderboard row

2026

PulseAnchor

PulseAnchor: Personalizing PPG Foundation Models for Cuffless Blood Pressure Estimation

7.84↓

Cuffless blood pressure · aurora-bp · Mean absolute error (mmHg); cuffless BP

uses extra training data

Leaderboard row

2026

Hybrid features + personalized calibration

In Personalized Calibration and Hybrid Feature Fusion for Continuous Blood Pressure Estimation Using PPG and ECG Signals

7.85↓

Cuffless blood pressure · BP-UCI · MAE (mmHg)

not stated in the abstractwithin-datasetfeature-based model with base-BP calibrationSBP

Abstract (Results)

2026

Awake–sleep PPG temporal indices + SVM

Hypertension screening via awake–sleep differences in photoplethysmogram signals

0.846↑

Cuffless blood pressure · cuhk-bp-htn · Accuracy

Protocol not structured

Leaderboard row

2026

3M-BPNet

Utilizing Multi-PPG-Sensor Site Information in a Localized Wrist Area for Improving Cuffless Blood Pressure Estimation

4.84↓

Cuffless blood pressure · liu-wrist-9ch-array-121 · Mean absolute error (mmHg); cuffless BP

subjects overlap

Leaderboard row

2026

McBP-Net

Dynamic Beat-to-Beat Measurements of Blood Pressure Using Multimodal Physiological Signals and a Hybrid CNN-LSTM Model

4.19↓

Cuffless blood pressure · mcbp-cold-pressor-23 · Mean absolute error (mmHg); cuffless BP

Protocol not structured

Leaderboard row

2025

FiLMamba

In FiLMamba: A FiLM-Conditioned Mamba Architecture with an Energy-Efficient FPGA Accelerator for Generalizable Cuffless Blood Pressure Estimation

6.45↓

Cuffless blood pressure · MIMIC-BP (1,524 patients) · SD of error (mmHg)

not stated in the abstractwithin-datasetFiLM-conditioned Mamba, end-to-endSBP

Abstract (Results)

2026

FiLMamba

FiLMamba: A FiLM-Conditioned Mamba Architecture with an Energy-Efficient FPGA Accelerator for Generalizable Cuffless Blood Pressure Estimation

6.45↓

Cuffless blood pressure · MIMIC-BP · Standard deviation of error (mmHg); cuffless BP

Protocol not structured

Leaderboard row

2026

Direct PPG→BP pipeline

In Blood Pressure Estimation from PPG: A Comparative Study of Direct and ECG-Mediated Deep Learning Pipelines

4.82↓

Cuffless blood pressure · MIMIC-III waveform · MAE (mmHg)

segment-level train/validation/test split (1.74M / 218k / 217k segments; patient separation not stated)within-datasetend-to-end training, no subject calibrationSBP

Table V

2026

Multiresolution U-Net @62.5 Hz

Practical optimization of deep learning models for cuffless blood pressure estimation from photoplethysmography

5.24↓

Cuffless blood pressure · mimic-waveform-205k · Mean absolute error (mmHg); cuffless BP

Protocol not structured

Leaderboard row

2026

Soft, Multi-Wavelength Photoplethysmography Enab

Soft, Multi-Wavelength Photoplethysmography Enables Reliable Neonatal Blood Pressure Monitoring Via Error Stratification

9.67↓

Cuffless blood pressure · neo-bp · Mean absolute error (mmHg); cuffless BP

Protocol not structured

Leaderboard row

2026

McBP-Net

In Dynamic Beat-to-Beat Measurements of Blood Pressure Using Multimodal Physiological Signals and a Hybrid CNN-LSTM Model

4.19↓

Cuffless blood pressure · Own cold-pressor recordings (23 subjects) · MAE (mmHg)

not stated in the abstractwithin-datasetmultimodal CNN-LSTM, beat-to-beatSBP

Abstract (Results)

2025

Windkessel-regularized model

In Cuffless Blood Pressure Estimation From Wearable Photoplethysmography and Windkessel Model-Regularized Learning Approach

4.10↓

Cuffless blood pressure · Own finger-wearable recordings (10 healthy subjects) · MAE (mmHg)

not stated in the abstractwithin-datasetphysics-regularized deep modelSBP

Abstract (Results)

2026

3M-BPNet

In Utilizing Multi-PPG-Sensor Site Information in a Localized Wrist Area for Improving Cuffless Blood Pressure Estimation

4.84↓

Cuffless blood pressure · Own wrist 9-channel PPG array (121 subjects) · MAE (mmHg)

not stated in the abstractwithin-datasetmulti-site PPG modelSBP

Abstract (Results)

2026

TEX-MEX 0.56M PPG FM

TEX-MEX

0.7347↑

Cuffless blood pressure · PPG-BP Database (Guilin / Liang 2018) · Area under ROC

uses extra training data

Leaderboard row

2026

Chowdhury GPR + ReliefF

Estimating Blood Pressure from the Photoplethysmogram Signal and Demographic Features Using Machine Learning Techniques

3.02↓

Cuffless blood pressure · PPG-BP Database (Guilin / Liang 2018) · Mean absolute error (mmHg); cuffless BP

subjects overlap

Leaderboard row

2020

TF-MambaNet

TF-MambaNet: A Temporal and Frequency Fused Bidirectional Mamba Architecture for PPG Foundation Model

12.89↓

Cuffless blood pressure · PPG-BP Database (Guilin / Liang 2018) · Mean absolute error (mmHg); cuffless BP

uses extra training data

Leaderboard row

2026

AnyPPG

PPG-Bench: A Unified Benchmark for PPG Foundation Models

13.32↓

Cuffless blood pressure · PPG-BP Database (Guilin / Liang 2018) · Mean absolute error (mmHg); cuffless BP

subject-disjointuses extra training data

Leaderboard row

2026

TEX-MEX 0.56M PPG FM

TEX-MEX

14.007↓

Cuffless blood pressure · PPG-BP Database (Guilin / Liang 2018) · Mean absolute error (mmHg); cuffless BP

uses extra training data

Leaderboard row

2026

Pulse-PPG

In Pulse-PPG: An Open-Source Field-Trained PPG Foundation Model for Wearable Applications Across Lab and Field Settings

8.878↓

Cuffless blood pressure · PPG-BP · MAE (mmHg)

subject-wisewithin-datasetlinear probe on frozen embeddingsdiastolic BP

Table 6

2025

PaPaGei

Baseline in Pulse-PPG: An Open-Source Field-Trained PPG Foundation Model for Wearable Applications Across Lab and Field Settings

9.246↓

Cuffless blood pressure · PPG-BP · MAE (mmHg)

subject-wisewithin-datasetlinear probe on frozen embeddingsdiastolic BP

Table 6 (as reported)

2024

MOMENT

Baseline in Pulse-PPG: An Open-Source Field-Trained PPG Foundation Model for Wearable Applications Across Lab and Field Settings

9.447↓

Cuffless blood pressure · PPG-BP · MAE (mmHg)

subject-wisewithin-datasetlinear probe on frozen embeddingsdiastolic BP

Table 6 (as reported)

2024

Chronos

Baseline in Pulse-PPG: An Open-Source Field-Trained PPG Foundation Model for Wearable Applications Across Lab and Field Settings

9.588↓

Cuffless blood pressure · PPG-BP · MAE (mmHg)

subject-wisewithin-datasetlinear probe on frozen embeddingsdiastolic BP

Table 6 (as reported)

2024

Dilated-residual attention U-Net

In Calibration-Based Cuffless Blood Pressure Estimation Using a Dilated-Residual Attention U-Net

2.03↓

Cuffless blood pressure · PulseDB (calibration-based) · MAE (mmHg)

calibration-based subset (same subjects in training and test)within-datasetend-to-end, PPG+ECG inputDBP

Abstract (Results)

2026

XResNet1d101

In Generalizable deep learning for photoplethysmography-based blood pressure estimation — A benchmarking study

13.9↓

Cuffless blood pressure · PulseDB (in-distribution, no calibration) · MAE (mmHg)

in-distribution test without subject-specific calibrationwithin-datasetend-to-endSBP

Abstract (Results)

2025

GPT-PPG

GPT-PPG: a GPT-based foundation model for photoplethysmography signals

8.12↓

Cuffless blood pressure · PulseDB v2.0 · Mean absolute error (mmHg); cuffless BP

subject-disjointuses extra training data

Leaderboard row

2025

Dilated-residual attention U-Net

Calibration-Based Cuffless Blood Pressure Estimation Using a Dilated-Residual Attention U-Net

3.81↓

Cuffless blood pressure · pulsedb-calbased · Mean absolute error (mmHg); cuffless BP

Protocol not structured

Leaderboard row

2026

PPG-FusionNet

PPG-FusionNet: A Dual-Branch Neural Architecture for Cuffless Blood Pressure Estimation from Photoplethysmography

7.46↓

Cuffless blood pressure · pulsedb-calbased · Mean absolute error (mmHg); cuffless BP

Protocol not structured

Leaderboard row

2026

ECG-PPG JEPA 345M

Bridging ECG and PPG: Latent-Space Prediction for Robust Physiological Analysis

8.23↓

Cuffless blood pressure · pulsedb-vital · Mean absolute error (mmHg); cuffless BP

uses extra training data

Leaderboard row

2026

Few-shot on frozen FM embeddings

Data Efficient Cuffless Blood Pressure Classification from PPG using Foundation Model Embeddings and Few-Shot Learning

0.824↑

Cuffless blood pressure · pulsedb-vitaldb-aha6 · Accuracy

uses extra training data

Leaderboard row

2026

DPIKAN

DPIKAN: Decomposed Physics-Informed KANs for Physiological Signal Representation Learning

5.41↓

Cuffless blood pressure · ring-hgcpt · Standard deviation of error (mmHg); cuffless BP

subject-disjoint

Leaderboard row

2026

ExpertoRhythm

In ExpertoRhythm: Morphology-Aware Learning for Waveform Reconstruction and Cuffless Blood Pressure Estimation from Single-Channel PPG

2.46↓

Cuffless blood pressure · UCI cuff-less BP (942 subjects) · MAE (mmHg)

subject-block split: fixed 25% held-out test, 5-fold CV on the rest for model selectionwithin-datasetPPG→ABP waveform reconstruction, end-to-endSBP

Table 3

2026

SimCLR

Baseline in AnyPPG: An ECG-Guided PPG Foundation Model Trained on Over 100,000 Hours of Recordings for Holistic Health Profiling

7.14↓

Cuffless blood pressure · UCI-BP · MAE (mmHg)

as defined per dataset in the paperwithin-datasetlinear probe on frozen embeddingsdiastolic BP

Table 3 (as reported)

2020

AnyPPG

In AnyPPG: An ECG-Guided PPG Foundation Model Trained on Over 100,000 Hours of Recordings for Holistic Health Profiling

7.14↓

Cuffless blood pressure · UCI-BP · MAE (mmHg)

as defined per dataset in the paperwithin-datasetlinear probe on frozen embeddingsdiastolic BP

Table 3

2025

BYOL

Baseline in AnyPPG: An ECG-Guided PPG Foundation Model Trained on Over 100,000 Hours of Recordings for Holistic Health Profiling

7.31↓

Cuffless blood pressure · UCI-BP · MAE (mmHg)

as defined per dataset in the paperwithin-datasetlinear probe on frozen embeddingsdiastolic BP

Table 3 (as reported)

2020

MOMENT

Baseline in AnyPPG: An ECG-Guided PPG Foundation Model Trained on Over 100,000 Hours of Recordings for Holistic Health Profiling

7.53↓

Cuffless blood pressure · UCI-BP · MAE (mmHg)

as defined per dataset in the paperwithin-datasetlinear probe on frozen embeddingsdiastolic BP

Table 3 (as reported)

2024

PulsePPG

Baseline in AnyPPG: An ECG-Guided PPG Foundation Model Trained on Over 100,000 Hours of Recordings for Holistic Health Profiling

7.57↓

Cuffless blood pressure · UCI-BP · MAE (mmHg)

as defined per dataset in the paperwithin-datasetlinear probe on frozen embeddingsdiastolic BP

Table 3 (as reported)

2025

ExpertoRhythm

ExpertoRhythm: Morphology-Aware Learning for Waveform Reconstruction and Cuffless Blood Pressure Estimation from Single-Channel PPG

2.46↓

Cuffless blood pressure · uci-bp · Mean absolute error (mmHg); cuffless BP

Protocol not structured

Leaderboard row

2026

Showing 51–100 of 433 results

Load 50 more results

Recent changes

Oct 27, 2026
New entry: A cross-subject stress recognition method…, 0.9158
Stress detection (acute / prolonged / affect): WESAD: accuracy
Oct 27, 2026
New entry: SPAFT, 5.066
Heart rate estimation: PPG-DaLiA: MAE (BPM)
Oct 27, 2026
New entry: SPAFT, 8.345
Cuffless blood pressure estimation: uci-bp: MAE (mmHg)
Oct 25, 2026
Leaderboard added with 1 result
Cuffless blood pressure estimation: neo-bp: MAE (mmHg)
Oct 23, 2026
New entry: PC-AVCT wearable-state model, 7.697
Cuffless blood pressure estimation: aurora-bp: MAE (mmHg)
Oct 23, 2026
New entry: PPG-FusionNet, 7.46
Cuffless blood pressure estimation: pulsedb-calbased: MAE (mmHg)

From the hub's dated sync history: new results as they are added, and SOTA stars as they are awarded or move.

How SOTA is decided

marks the best value for one task, dataset and metric, using the metric's direction (higher or lower is better).
Results are compared only under the same evaluation protocol: split type, within- or cross-dataset testing, filtering, inputs and label set. Different protocols stay separate and are never ranked together.
Only results with a table, figure or section reference compete, and a star needs at least two competing methods. Ties go to the earlier publication.
Stars are recomputed whenever the research hub syncs, and every change is logged.

Downloads of every result (JSON and CSV) are free for members. Sign in to download.

Results as reported by the authors; not independently reproduced. Leaderboard rows without a structured protocol are listed but don't receive stars. See also the leaderboards.

Cite or reuse this

CC BY 4.0ChatPPG compiled this data. You can copy, adapt and republish it, including in articles, slides, papers and AI tools, as long as you credit ChatPPG and link to this page. Please also cite the original studies for any number you use.

Attribution

Source: ChatPPG, Wearable AI results explorer, https://chatppg.com/research/results (CC BY 4.0)

HTML link

Source: <a href="https://chatppg.com/research/results">ChatPPG, Wearable AI results explorer</a> (<a href="https://creativecommons.org/licenses/by/4.0/" rel="license">CC BY 4.0</a>)