Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

EcoScaleNet: A Lightweight Multi Kernel Network for Long Sequence 12 lead ECG Classification

Dong-Hyeon Kang, Ju-Hyeon Nam, Sang-Chul Lee

TL;DR

This work tackles long‑sequence 12‑lead ECG classification by addressing the inefficiency of Omni‑Scale Convolution (OS‑CNN), which uses a large prime‑kernel set to cover all receptive fields but incurs prohibitive compute. EcoScale‑Net replaces the exhaustive, costly design with a hierarchical framework that caps kernel sizes at deeper stages and introduces dual 1×1 bottlenecks around OS‑Conv blocks, preserving full RF coverage while enabling deeper/wider models. On CODE‑15% ECG data, EcoScale‑Net achieves a 90% reduction in parameters and 99% reduction in FLOPs, while improving macro‑averaged F1 by 2.4% and delivering state‑of‑the‑art accuracy for long‑sequence ECG classification, suitable for real‑time deployment on commodity hardware. The approach offers a Pareto‑optimal balance between accuracy and efficiency and is accompanied by publicly available code for reproducibility and extension to other biosignals.

Abstract

Accurate interpretation of 12 lead electrocardiograms (ECGs) is critical for early detection of cardiac abnormalities, yet manual reading is error prone and existing CNN based classifiers struggle to choose receptive field sizes that generalize to the long sequences typical of ECGs. Omni Scale CNN (OS CNN) addresses this by enumerating prime sized kernels inspired by Goldbach conjecture to cover every scale, but its exhaustive design explodes computational cost and blocks deeper, wider models. We present Efficient Convolutional Omni Scale Network (EcoScale-Net), a hierarchical variant that retains full receptive field coverage while eliminating redundancy. At each stage, the maximum kernel length is capped to the scale still required after down sampling, and bottleneck convolutions inserted before and after every Omni Scale block curtail channel growth and fuse multi scale features. On the large scale CODE 15% ECG dataset, EcoScaleNet reduces parameters by 90% and FLOPs by 99% compared with OS CNN, while raising macro averaged F1 score by 2.4%. These results demonstrate that EcoScaleNet delivers SOTA accuracy for long sequence ECG classification at a fraction of the computational cost, enabling real time deployment on commodity hardware. Our EcoScaleNet code is available in GitHub Link.

EcoScaleNet: A Lightweight Multi Kernel Network for Long Sequence 12 lead ECG Classification

TL;DR

This work tackles long‑sequence 12‑lead ECG classification by addressing the inefficiency of Omni‑Scale Convolution (OS‑CNN), which uses a large prime‑kernel set to cover all receptive fields but incurs prohibitive compute. EcoScale‑Net replaces the exhaustive, costly design with a hierarchical framework that caps kernel sizes at deeper stages and introduces dual 1×1 bottlenecks around OS‑Conv blocks, preserving full RF coverage while enabling deeper/wider models. On CODE‑15% ECG data, EcoScale‑Net achieves a 90% reduction in parameters and 99% reduction in FLOPs, while improving macro‑averaged F1 by 2.4% and delivering state‑of‑the‑art accuracy for long‑sequence ECG classification, suitable for real‑time deployment on commodity hardware. The approach offers a Pareto‑optimal balance between accuracy and efficiency and is accompanied by publicly available code for reproducibility and extension to other biosignals.

Abstract

Accurate interpretation of 12 lead electrocardiograms (ECGs) is critical for early detection of cardiac abnormalities, yet manual reading is error prone and existing CNN based classifiers struggle to choose receptive field sizes that generalize to the long sequences typical of ECGs. Omni Scale CNN (OS CNN) addresses this by enumerating prime sized kernels inspired by Goldbach conjecture to cover every scale, but its exhaustive design explodes computational cost and blocks deeper, wider models. We present Efficient Convolutional Omni Scale Network (EcoScale-Net), a hierarchical variant that retains full receptive field coverage while eliminating redundancy. At each stage, the maximum kernel length is capped to the scale still required after down sampling, and bottleneck convolutions inserted before and after every Omni Scale block curtail channel growth and fuse multi scale features. On the large scale CODE 15% ECG dataset, EcoScaleNet reduces parameters by 90% and FLOPs by 99% compared with OS CNN, while raising macro averaged F1 score by 2.4%. These results demonstrate that EcoScaleNet delivers SOTA accuracy for long sequence ECG classification at a fraction of the computational cost, enabling real time deployment on commodity hardware. Our EcoScaleNet code is available in GitHub Link.

Paper Structure

This paper contains 16 sections, 9 equations, 1 figure, 3 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Method
  3. Preliminary: Omni‑Scale Convolution
  4. Kernel set.
  5. Receptive‑field coverage.
  6. Computational bottleneck.
  7. Overall Architecture
  8. EcoScale Residual Block
  9. Hierarchical RF cascading.
  10. Channel‑efficient fusion.
  11. Complexity Analysis and Receptive Field of EcoScale-Net
  12. Experimental Results
  13. Experimental Settings and Implementation Details
  14. Results Analysis
  15. Ablation Study
  16. ...and 1 more sections

Figures (1)

  • Figure 1: (a) Overview of EcoScale‑Net. (b) EcoScale Block which is main module for proposed approach.