MFConvTr: Multi-Frequency Convolutional Transformer for Fetal Arrhythmia Detection in Non-Invasive fECG

Deva Satya Sriram Chintapenta ∗, Aman Verma ∗, Saikat Majumder · 2025

Read MFConvTr: Multi-Frequency Convolutional Transformer for Fetal Arrhythmia Detection in Non-Invasive fECG by Deva Satya Sriram Chintapenta ∗, Aman Verma ∗, Saikat Majumder (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-16649.txt.

Description

Chintapenta, Verma, and Majumder present a deep learning model that spots irregular heartbeats in unborn babies from belly-surface recordings, in this 2025 arXiv paper. Such recordings mix the mother's and baby's signals, making them hard to read, and classic marker-based methods fail on their tangled shapes. Odd rhythms show up in both the timing and the frequency makeup of the trace.

Their model, MFConvTr, mixes convolutional layers that read several frequency bands at once with a transformer that tracks long-range patterns. Splitting the signal into channels and filtering each at a different scale captures the multi-frequency clues cheaply. The model reached top scores on its tests while using very few parameters. The authors also ran breakdown tests to show which parts carry the gains.

Catalog record

CREATORDeva Satya Sriram Chintapenta ∗, Aman Verma ∗, Saikat Majumder
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE18 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-16649

Browse the full library · About the Tower of Babel