EURASIP Journal on Advances in Signal Processing

Table 6 Accuracy (%) of seal calls recognition by the ResNet18 [29] model

From: Seal call recognition based on general regression neural network using Mel-frequency cepstrum coefficient features

SNR/dB	MFCC				LOFAR spectrum
SNR/dB	A	B	C	Total	A	B	C	Total
20	98.32	93.80	92.82	94.98	94.25	90.19	89.52	91.32
15	96.92	93.76	91.38	94.02	93.85	89.09	89.61	90.85
10	94.21	93.22	90.31	92.58	88.03	85.31	83.71	85.69
5	91.55	90.15	85.06	88.92	85.19	83.22	80.38	82.93
0	85.59	82.92	80.28	82.56	74.52	74.41	71.21	73.38

Back to article page