EURASIP Journal on Advances in Signal Processing

Table 7 Accuracy (%) of seal calls recognition by the SVM [26] 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	94.51	97.31	92.52	94.78	91.85	88.02	89.17	89.68
15	92.10	95.25	90.33	92.56	89.36	86.71	85.31	87.12
10	91.10	93.35	89.66	91.37	84.64	86.47	81.37	84.16
5	88.05	85.39	83.02	85.48	81.66	81.85	79.91	81.14
0	81.65	82.00	81.63	81.76	74.02	70.92	68.03	70.99

Back to article page