Skip to main content

Table 3 Reduction in over- and undersensing. Detection optimization for the Rel-En and AMM algorithms using both linear and non-linear weight functions. The values in brackets refer to the relative variations in error rates with the optimization

From: Evaluation and optimization of novel extraction algorithms for the automatic detection of atrial activations recorded within the pulmonary veins during atrial fibrillation

 

False negative

(% change with optimization)

False positive

(% change with optimization)

Total error

(% change with optimization)

Raw Rel-En

0.76%

0.12%

0.88%

 Linear

0.40% (− 47.4%)

0.07% (− 37.9%)

0.47% (− 46.1%)

 Non-linear

0.24% (− 68.4%)

0.04% (− 70.6%)

0.28% (− 68.7%)

Raw AMM

2.65%

0.84%

3.49%

 Linear

2.14% (− 19.2%)

1.17% (+ 39.3%)

3.31% (− 5.2%)

 Non-linear

2.25% (− 15.1%)

1.08% (+ 28.6%)

3.33% (− 4.6%)