ECG-IMN: Interpretable Mesomorphic Neural Networks for 12-Lead Electrocardiogram Interpretation
Vajira Thambawita 1 , Jonas L. Isaksen 2 , Jørgen K. Kanters 2 , Hugo L. Hammer 3 , Pål Halvorsen 1 · 2026
Read ECG-IMN: Interpretable Mesomorphic Neural Networks for 12-Lead Electrocardiogram Interpretation by Vajira Thambawita 1 , Jonas L. Isaksen 2 , Jørgen K. Kanters 2 , Hugo L. Hammer 3 , Pål Halvorsen 1 (2026) 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-2602-09566.txt.
Description
Deep learning matches experts on auto ECG readings, yet black-box models stall clinic uptake: doctors need to see which body traits drive each call. Vajira Thambawita and four co-authors propose ECG-IMN, a 2026 clear network for high-detail twelve-lead ECG sorting. Standard explain tools guess after the fact, with unstable, costly, and sometimes false credit.
ECG-IMN instead works as a hyper-net: a deep filter backbone makes settings of a strictly linear model fitted to each input sample. Choice logic is thus clear in math, and the made weights serve as exact, high-detail credit maps. A bridge decoder maps latent traits to per-sample weights, spotting diseased evidence such as ST rise and T-wave flip in time and across leads. On the PTB-XL dataset, scores rival black-box rivals while reasons stay faithful and own to each case.
Catalog record
| CREATOR | Vajira Thambawita 1 , Jonas L. Isaksen 2 , Jørgen K. Kanters 2 , Hugo L. Hammer 3 , Pål Halvorsen 1 |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 28 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-09566