Guideline-grounded global evaluation of 13 gradient XAI methods on ECG classifiers shows amplitude-linked methods miss low-amplitude diagnostic regions, while SIGN-adjusted variants maintain worst-case coverage on all four pathologies.
R.et al.Discovering and visualizing disease-specific electrocardiogram features using deep learning.Circ
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CY 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Beyond Local Inspection: Global, Guideline-Grounded Evaluation of Post-hoc XAI Methods for ECG Classification
Guideline-grounded global evaluation of 13 gradient XAI methods on ECG classifiers shows amplitude-linked methods miss low-amplitude diagnostic regions, while SIGN-adjusted variants maintain worst-case coverage on all four pathologies.