Skip to content
DraconiticResearch
Search Sources API
203,948 studies31 sourcesUpdated 8 min ago
Peer-reviewedCC BY 4.0PLOS One

STCP-Net: A few-shot learning approach based on spectral-temporal convolution and an improved prototypical network for ECG classification

  • Jia Li1
  • Dainan Gan2
  • Xiaolin Zhan3
  • Wei Yao4
  • Bin Zou5
  • Bin Zhang6
  • Dong Li
  • Dan Li
  • Lin Li
6 affiliations
  1. 1School of Electronics and Information Engineering, Wuyi University, Jiangmen, China
  2. 2School of Mathematics and Computational Science, Wuyi University, Jiangmen, China
  3. 3Department of Cardiovascular Disease and Clinical Experimental Center, Jiangmen Central Hospital, Jiangmen, China
  4. 4Department of Intensive Care Unit, Jiangmen Central Hospital, Jiangmen, China
  5. 5Department of Electrocardiogram, Jiangmen Central Hospital, Jiangmen, China
  6. 6Department of Rehabilitation, Shengjing Hospital of China Medical University, Shenyang, China
Open at sourcePDFDOICite

Abstract

To address the heavy reliance of deep learning models on manually labeled data in ECG classification, we need to tackle two major limitations. First, sample sizes are often limited in clinical settings, making data annotation costly and impractical. Second, existing few-shot learning methods still fall short in two key areas: 1) Existing strategies pay insufficient attention to rare arrhythmia types. This makes it difficult to effectively identify rare clinical cases. 2) Single feature representation fails to fully integrate the global frequency, local time-frequency, and semantic information of ECG signals. This results in incomplete feature representation. To overcome these challenges, we propose a novel few-shot learning method for ECG classification. In terms of feature extraction, a spectral-temporal convolution feature extractor is designed to precisely focus on the key information of ECG by efficiently integrating the global frequency, local time-frequency features and semantic information of the signal. Furthermore, with regard to classification, an improved prototypical network based on cross-class bias diminishing is developed to enhance the recognition performance for rare arrhythmia type samples. The experimental results show that the proposed method achieves significant classification performance across multiple public datasets, especially on Chapman, it attains an average accuracy exceeding 90% in the 2-way 1-shot and 2-way 5-shot tasks.

Related

  1. KT-CBHNET: a spatio-temporal deep learning framework for interpretable ECG signal classification for cardiovascular disease prediction
    Krantee Singh, Vasudev Dehalwar
    Scientific ReportsPeer-reviewedCC BY-NC-ND 4.0Full textCardiovascular medicineCrossref
  2. Temporal deep learning framework for multi condition fault classification of three-phase induction motors
    R. Sooraj, Senthil Kumar Ramu, R. Sitharthan +2
    Scientific ReportsPeer-reviewedCC BYFull textGeneral medicineDOAJ
  3. FedRAPT: federated representation-aligned prototypical contrastive learning
    Su-Bin Seo, Jong-Hoon Kim, Chang-Sun Shin +3
    Scientific ReportsPeer-reviewedCC BYFull textGeneral medicineDOAJ
  4. Machine Learning with Social Determinants of Health for Cross-Sectional Diabetes Classification: Development and Temporal Validation.
    Woon-Man Kung, Hsien Chang Yeh, Abel Po-Hao Huang +3
    International journal of medical sciencesPeer-reviewedCC BYFull textGeneral medicineEurope PMC
  5. Multi-scale convolution with efficient channel attention improves benign–malignant classification of spinal tumors on T2-weighted fat-suppressed MRI
    Zhiwei Ma, Xingyu Duan, Xiaorong Ma +3
    Frontiers in OncologyPeer-reviewedCC BY 4.0OncologyFrontiersCrossref

Cite

BibTeXRIS
BibTeX
@article{li2026stcpnet,
  title = {{STCP-Net: A few-shot learning approach based on spectral-temporal convolution and an improved prototypical network for ECG classification}},
  author = {Jia Li and Dainan Gan and Xiaolin Zhan and Wei Yao and Bin Zou and Bin Zhang and Dong Li and Dan Li and Lin Li},
  journal = {PLOS One},
  year = {2026},
  month = oct,
  doi = {10.1371/journal.pone.0357954},
  url = {https://doi.org/10.1371/journal.pone.0357954},
  note = {CC BY 4.0},
}
RIS
TY  - JOUR
TI  - STCP-Net: A few-shot learning approach based on spectral-temporal convolution and an improved prototypical network for ECG classification
AU  - Jia Li
AU  - Dainan Gan
AU  - Xiaolin Zhan
AU  - Wei Yao
AU  - Bin Zou
AU  - Bin Zhang
AU  - Dong Li
AU  - Dan Li
AU  - Lin Li
PY  - 2026
DA  - 2026/10/15
JO  - PLOS One
DO  - 10.1371/journal.pone.0357954
UR  - https://doi.org/10.1371/journal.pone.0357954
AB  - To address the heavy reliance of deep learning models on manually labeled data in ECG classification, we need to tackle two major limitations. First, sample sizes are often limited in clinical settings, making data annotation costly and impractical. Second, existing few-shot learning methods still fall short in two key areas: 1) Existing strategies pay insufficient attention to rare arrhythmia types. This makes it difficult to effectively identify rare clinical cases. 2) Single feature representation fails to fully integrate the global frequency, local time-frequency, and semantic information of ECG signals. This results in incomplete feature representation. To overcome these challenges, we propose a novel few-shot learning method for ECG classification. In terms of feature extraction, a spectral-temporal convolution feature extractor is designed to precisely focus on the key information of ECG by efficiently integrating the global frequency, local time-frequency features and semantic information of the signal. Furthermore, with regard to classification, an improved prototypical network based on cross-class bias diminishing is developed to enhance the recognition performance for rare arrhythmia type samples. The experimental results show that the proposed method achieves significant classification performance across multiple public datasets, especially on Chapman, it attains an average accuracy exceeding 90% in the 2-way 1-shot and 2-way 5-shot tasks.
SN  - 1932-6203
ER  -