Cardiac disease classifier using Digital Twin

An cardiac disease classifier in order to diagnose cardiac diseases with artificial intelligence. 

We used both synthetic and experimental data for training our machine learning model.  Various features were extracted from synthetic and experimental hemo dynamic/acoustic factors for different cardiac diseases.

Total accuracy of the model was 93.5% for classification of 10 common cardiac diseases with only 6 features.

All the Classified Common Cardiac Diseases

Normal

Normal condition and our reference.

HF

Complete heart failure

HFLv

Left ventricular heart failure

HFRv

Right ventricular heart failure

AVd

Accelerated atrioventricular conduction

LBBB

Left bundle branch block

RBBB

Right bundle branch block

ASD

Atrial septal defect

VSD

Ventricular septal defect

M1Leak

Mitral valve leak

T1Leak

Tricuspid valve leak

A2Leak

Aortic valve leak

P2Leak

Pulmonic valve leak

A2Leak

Aortic valve leak

A2Leak

Aortic valve leak

M1Sten

Mitral valve stenosis

T1Sten

Tricuspid valve stenosis

A2Sten

Aortic valve stenosis

P2Sten

Pulmonic valve stenosis

AVdDec

Decelerated atrioventricular conduction