Key result
Lightweight ECG classifiers show poor generalization, achieving only ~17% supraventricular recall on MIT-BIH testing.
Why the study?
Efficient microcontroller execution does not ensure reliable ECG arrhythmia classification across patients and databases, motivating a controlled benchmark evaluation.
In lightweight ECG arrhythmia classification on microcontrollers, performance is limited more by patient-specific variability and false-positive predictions than by numerical precision or hardware constraints.
Lightweight ECG classifiers may lack reliability for supraventricular detection; leaves open need for robust, patient-independent algorithms before clinical use.
Efficient microcontroller execution does not ensure reliable electrocardiogram classification across patients and databases. We conducted a controlled benchmark of three convolutional architectures under a corrected subject-disjoint MIT-BIH split, external INCART testing, Noise Stress Test Database artifacts, 16-bit floating-point (FP16) and full-integer 8-bit (INT8) conversion, and classifier-only invocation on an ESP32 without external RAM. Five fixed seeds were used. On the MIT-BIH test cohort (DS2), INT8 four-class macro-F1 was 0.474-0.498 and supraventricular (S) recall was 0.150-0.180. Mean S-class area under the receiver operating characteristic curve (AUROC) was below 0.5 in all six architecture-regime groups, although all corresponding record-cluster intervals included 0.5. Within-record ranking could remain high despite poor pooled ranking. On INCART, S-class ranking was stronger, with all six patient-cluster AUROC intervals above 0.5, but the representative mixed-noise depthwise-separable network (DS-CNN) achieved 0.797 recall with 0.074 precision. Including record 201, which shares a subject with test record 202, increased DS2 S-recall point estimates by 2.49-14.68 percentage points. In the primary architecture panel, FP16 and INT8 met the engineering tolerances in all 120 comparisons. DS-CNN required 9.49 ms per classifier invocation on the ESP32 using preprocessed, annotation-centered inputs. Acquisition, QRS detection, preprocessing, current, and energy were not measured. Patient-dependent S-score behavior and external false-positive predictions limited this panel more than numerical precision or microcontroller resources.
No takes yet. Share an insight, caveat, or question.
Park et al. (2026) studied Electrocardiogram arrhythmia classification. Lightweight convolutional architectures for ECG classification was evaluated on INT8 four-class macro-F1 and supraventricular (S) recall on the MIT-BIH test cohort. Evaluation of lightweight ECG classifiers showed poor generalization, with INT8 supraventricular recall of 0.150-0.180 on MIT-BIH, limited by patient-dependent behavior and false positives.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: