Beat Classification Task

Overview

In beat classification, we classify individual beats as either normal or abnormal. Abnormal beats can be further classified as being either premature or escape beats as well as originating from the atria, junction, or ventricles. The objective of beat classification is to detect and classify these abnormal heart beats directly from ECG signals.

---

Characteristics

	Atrial	Junctional	Ventricular
Premature	PAC P-wave: Different QRS: Narrow (normal) Aberrated: LBBB or RBBB	PJC P-wave: None / retrograde QRS: Narrow (normal) Compensatory SA Pause	PVC P-wave: None QRS: Wide (> 120 ms) Compensatory SA Pause
Escape	Atrial Escape	P-wave: Abnormal QRS: Narrow (normal) Ventricular rate: < 60 bpm Junctional Escape	P-wave: None QRS: Narrow (normal) Bradycardia (40-60 bpm) Ventricular Escape

---

Dataloaders

Dataloaders are available for the following datasets:

• Icentia11k
• PTB-XL

---

Pre-trained Models

The following table provides the latest performance and accuracy results for pre-trained beat models. Additional result details can be found in Model Zoo → Beat.

NAME	DATASET	FS	DURATION	# CLASSES	MODEL	PARAMS	FLOPS	METRIC
BC-2-EFF-SM	Icentia11k	100Hz	5s	2	EfficientNetV2	28K	1.8M	97.7% F1
BC-3-EFF-SM	Icentia11k	100Hz	5s	3	EfficientNetV2	41K	2.1M	92.0% F1

---

Target Classes

Below outlines the classes available for beat classification. When training a model, the number of classes, mapping, and names must be provided.

CLASS	LABELS
0	Normal
1	PAC
2	PVC
3	Noise

Class Mapping

Below is an example of a class mapping for a 3-class beat model. The class map keys are the original class labels and the values are the new class labels. Any class not included will be skipped.

JSONC example

{
    "num_classes": 3,
    "class_names": ["QRS", "PAC", "PVC"],
    "class_map": {
        "0": 0,  // Map Normal to QRS
        "1": 1,  // Map PAC to PAC
        "2": 2,  // Map PVC to PVC
    }
}

---