Noise-resistant network: a deep-learning method for face recognition under noise

EURASIP Journal on Image and Video Processing

Table 4 The average recognition rates of different methods on the AR database injected with uniform noise

Algorithm	Chi-square distance, p =				Histogram intersection, p =				Modified G-statistics, p =
Algorithm	0.10	0.20	0.40	0.70	0.10	0.20	0.40	0.70	0.10	0.20	0.40	0.70
FLBP	0.7932	0.7574	0.6017	0.5055	0.7889	0.7623	0.6080	0.5184	0.7801	0.7562	0.6198	0.4827
NRLBP	0.7999	0.7670	0.6244	0.5159	0.8018	0.7624	0.6205	0.5020	0.7991	0.7571	0.6282	0.4732
NRLBP+	0.8264	0.7747	0.6804	0.5465	0.8222	0.7843	0.6731	0.5263	0.8201	0.7769	0.6849	0.5233
NRLBP++	0.8313	0.7945	0.6936	0.5363	0.8226	0.7816	0.6812	0.5295	0.8291	0.7861	0.6889	0.5321
	Pearson correlation coefficient, p =				Euclidean distance, p =				Cosine distance, p =
	0.10	0.20	0.40	0.70	0.10	0.20	0.40	0.70	0.10	0.20	0.40	0.70
BN2	0.8489	0.8065	0.7172	0.6438	0.8391	0.8028	0.6990	0.6289	0.8462	0.8109	0.7205	0.6462
BN1	0.8496	0.8375	0.7531	0.6997	0.8478	0.8342	0.7421	0.7025	0.8515	0.8418	0.7565	0.7063
NR-Network	0.8687	0.8463	0.7974	0.7242	0.8643	0.8409	0.7884	0.7165	0.8795	0.8559	0.7985	0.7275