Few-Shot Prototype Head Adaptation for On-Device ECG Personalization on PSoC~6

Guilherme Silva; Pedro Silva; Gladston Moreira; Eduardo Luz

cs.LGOct 5, 2026

Few-Shot Prototype Head Adaptation for On-Device ECG Personalization on PSoC~6

Authors: Guilherme Silva, Pedro Silva, Gladston Moreira, Eduardo Luz

Organizations: Computing Department, Universidade Federal de São João del-Rei – São João del-Rei, Brazil · Computing Department, Federal University of Ouro Preto – Ouro Preto, Brazil

Abstract

Wearable and bedside electrocardiogram (ECG) monitors must adapt to patient-specific morphology to maintain arrhythmia detection accuracy across users, yet personalization is typically performed offline and cannot account for individual physiology, electrode placement, or recording drift. On-device adaptation by backpropagation is expensive for microcontroller-class medical devices because it requires an optimizer state, repeated backward passes through convolutional layers, and labeled arrhythmic beats that may not be available at deployment time. This letter proposes prototype-only head adaptation as a compact personalization primitive for TinyML ECG systems. A one-dimensional convolutional neural network (1-D CNN; 1,314 parameters and 72.6k multiply-accumulate operations per beat) is trained offline on the MIT-BIH Arrhythmia Database under an inter-patient protocol, frozen as a feature extractor, and exported to a PSoC 6 microcontroller. Patient-specific adaptation then reduces to computing closed-form class means in a 32-dimensional embedding space, requiring no convolutional backward pass, no iterative optimization, and only one forward pass per support beat. Prototype adaptation improves inter-patient macro-F1 from 0.635/0.639/0.646 to 0.731/0.771/0.797 at 1/5/10-shot, outperforming linear stochastic-gradient-descent (SGD) head fine-tuning at every shot count for the target tiny backbone. On-device replay over 18 one-shot episodes on a PSoC 6 Cortex-M4F matches the host macro-F1 for the prototype head (0.798), with 11.39 ms per beat, 5.2 KB flash, and 22.2 KB SRAM. A restricted variant that updates only the normal-class prototype from passively buffered sinus beats yields a consistent +0.05 macro-F1 gain, reducing the annotation burden during initial

Figures & tables

Figure 1: Prototype-only ECG personalisation pipeline. Offline: a compact 1-D CNN is trained on DS1 under an inter-patient split; the frozen backbone is exported as native-C arrays. On-device (PSoC 6): the backbone maps support beats to 32-D embeddings; per-class prototypes are computed as closed-form means. Query beats are classified by nearest-prototype distance—with no backward pass, no optimizer state, and no backbone writes.

Item	Flash	SRAM
Backbone weights	4.9 KB	–
Head/prototypes	0.3 KB	0.3 KB
Peak activation buffer	–	6.3 KB
Support buffer, $K=10$	–	15.6 KB
Subtotal	5.2 KB	22.2 KB
PSoC 6 budget	1024 KB	288 KB

Table 1: Model-centric memory budget of the deployable adaptation path on PSoC 6. Weights are stored as float32 constants in flash.

Method	$K$	DS2 F1	Device F1	Adapt. time
No adaptation	–	0.664 †	0.611	–
Linear SGD	1	0.671	0.682	24.49 ms
Prototype	1	0.731	0.798	22.83 ms
Prototype	5	0.771	–	114.03 ms
Prototype	10	0.797	–	228.04 ms
Normal-only proto	10	0.699	–	$\approx$ 10 ms ∗

Table 2: Accuracy and embedded cost for the target tiny backbone. Device F1 is computed on the capped 18-record 1-shot replay subset. Timing is measured with the Arm Data Watchpoint and Trace (DWT) cycle counter on the PSoC 6 CM4.

Figure 2: Aggregate DS2 results across shot counts. Left : macro-F1 for each method. Right : software-side adaptation time from the host reference implementation. The pre-adaptation curve rises with

K

because the eligibility filter excludes harder records, not due to any adaptation effect.

Figure 3: Restricted vs. full prototype adaptation on the common eligible DS2 subset. The restricted variant (only normal-beat support from the target patient) provides a consistent positive gain at a fraction of the cost of full prototype adaptation.

Figure 4: End-to-end calibration time on the PSoC 6 CM4 at 100 MHz vs. number of support shots

K

for prototype and linear SGD adaptation. Both methods scale linearly with

K

(one backbone forward pass per support beat); the gap between the two curves is due solely to the head-update rule. The 400 ms RR interval at 150 BPM is shown for clinical reference.

Variant	Params	MACs	Pre F1	1-shot	10-shot
Tiny	1,314	72.6k	0.664	Proto .731	Proto .797
Medium	2,618	157.0k	0.642	SGD .713	Proto .764
Large	4,930	274.2k	0.651	SGD .720	Proto .759

Table 3: Scaling study across backbone sizes. DS2 test F1 is reported before personalization. Few-shot columns show the best retuned head-adaptation method.

Explore similar work

May 21, 2026cs.LG

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

Assessing cognitive load continuously and at low latency would help adaptive human-computer interaction, but it remains hard because labeled data are scarce and models generalize poorly across subjects. Recent ECG foundation models, pre-trained on millions of clinical diagnostic ECG recordings, yet they do not apply directly to wearable devices when the sensor configuration and the task both differ. We present CogAdapt, a framework that adapts a clinical ECG foundation model to wearable cognitive load assessment. CogAdapt has two parts. LeadBridge is a learnable adapter that maps 3-lead wearable signals to a 12-lead-compatible representation. ProFine is a progressive fine-tuning strategy that unfreezes encoder layers in stages while limiting representational drift in the pre-trained model. On two public datasets (CLARE and CL-Drive) under leave-one-subject-out cross-validation, CogAdapt reaches macro-F1 of 0.626 and 0.768, improving over from-scratch baselines by 11.2 and 16.1 percentage points. The results show that a clinical ECG pretraining can support subject-independent cognitive load assessment from wearable sensors.

Amir Mousavi, Erfan Nourbakhsh, Mohammad Sadegh Sirjani +5

Department of Computer Science, College of AI, Cyber and Computing, The University of Texas at San Antonio · Department of Neuroscience, Developmental and Regenerative Biology, College of Sciences, The University of Texas at San Antonio · Department of Educational Psychology, College of Education and Human Development, The University of Texas at San Antonio

Electrocardiogram Foundation Models Wearable Physiology Cognitive Load Cogadapt

Ask PDF

Apr 25, 2026cs.LG

Domain-Adapted Fine-Tuning of ECG Foundation Models for Multi-Label Structural Heart Disease Screening

Transthoracic echocardiography is the reference standard for confirming structural heart disease (SHD), but first-line screening is limited by cost, workflow burden, and specialist availability. We evaluated whether open pretrained electrocardiogram (ECG) foundation models can support echo-confirmed multi-label SHD detection using the public EchoNext Mini-Model benchmark. Six echocardiography-derived abnormalities were targeted: reduced left ventricular ejection fraction, increased left ventricular wall thickness, aortic stenosis, mitral regurgitation, tricuspid regurgitation, and right ventricular systolic dysfunction. Under a common pipeline, we compared engineered ECG features with gradient boosting, end-to-end waveform learning from scratch, and transfer from open ECG foundation models. We then applied in-domain self-supervised adaptation of an ECG foundation model (ECG-FM) on EchoNext waveforms followed by selective supervised fine-tuning, and evaluated trade-offs between discrimination and adaptation cost. Adapted ECG-FM models achieved the best overall performance: peak macro-AUROC 0.8509 and macro-AUPRC 0.4297, while a parameter-efficient operating point preserved AUROC (0.8501) and attained the highest fixed-threshold macro-F1 0.3691. Late fusion with covariates did not improve threshold-independent discrimination, and evaluated LoRA, alternative backbones, and mixture-of-foundations strategies did not surpass the best adapted single-backbone models. These results indicate that for ECG-based case finding and echocardiography triage, combining target-domain self-supervised adaptation with selective supervised updating of a pretrained ECG backbone is the most effective transfer strategy.

Duc N. Do, Minh N. Do, Dang Nguyen +14

1PASSIO Laboratory, North Carolina A&T State University, Greensboro, NC, USA · Department of Computer Science, University of Manitoba, Winnipeg, MB, Canada · 3Harvard T.H. Chan School of Public Health, Harvard University, Boston, MA, USA +4

Electrocardiogram Foundation Models Cardiac Domain-Adapted Diffusion

Ask PDF

Jun 1, 2026cs.LG

ArrythML: An Autoencoder-Based TinyML Approach for On-Device Arrhythmia Detection on Resource-Constrained Embedded Systems

Our work presents a method for ECG segmentation and arrhythmia detection using Tiny Machine Learning (TinyML) models for real-time, on-device inference on resource-constrained embedded systems. We develop INT8 quantized autoencoder-based TinyML models with minimal layers and parameters for embedded deployment. These models are evaluated using a custom dataset derived from the MIT-BIH Arrhythmia Database and validated in both PC-based simulations and on-device environments. For the evaluations, over 95,000 ECG segments are processed on an ESP32-S3 microcontroller running the TensorFlow Lite Micro runtime. Post-evaluation, detailed analysis, including annotation-wise and record-wise failure analysis, is conducted to characterize model behavior across diverse ECG morphologies and rhythm patterns and to explain missed detections. In several cases, apparent misclassifications may correspond to early or subtle anomaly patterns labeled as normal in the reference annotations, highlighting the model's sensitivity. A refined evaluation by filtering out ambiguous cases in the dataset shows that the best-performing DNN-based autoencoder achieves a recall of 84%, an F1-score of 79%, a model size of approximately 180 KB, and an inference latency of 9 ms on-device. These results demonstrate the feasibility of low-power, privacy-preserving embedded wearable systems capable of performing accurate arrhythmia detection entirely on-device.

Nagarajan S, Kurian Polachan

International Institute of Information Technology, Bangalore, India

Arrhythmia Classification Electrocardiography Wearable Devices Autoencoder Architectures

Ask PDF