Assessment of Electrocardiogram Rhythms by GoogLeNet Deep Neural Network Architecture

Table 4

Accuracy of arrhythmia classification by the expanded kernel model with increasing number of filters.

Number of filters	ECG segments (%)					Mean accuracy (%)
	3	4	5	6	7
1	95.1	94.4	92.7	92.4	94.1	93.7
3	95.6	95.1	95.6	93.8	94.5	94.9
5	96.4	95.5	95.1	94.8	94.4	95.2
7	96	95.8	95.6	94.7	94.4	95.3
9	96.2	96.1	95.9	94.8	95.3	95.7
11	96.1	95.7	96	95.6	94.8	95.7
13	96.5	96.1	95.5	95.4	95.1	95.7
15	96.4	96.3	96.1	95.5	94.7	95.8
17	96.7	96.1	96.3	95.6	95.6	96.0
19	96.5	96	96.2	95.6	95.1	95.9
Mean accuracy (%)	96.2	95.7	95.5	94.8	94.8