Table 3 Comparative analysis of precision, recall and F-measure for existing RVM and proposed NSVM-CNN

From: An effective image retrieval based on optimized genetic algorithm utilized a novel SVM-based convolutional neural network classifier

No of image retrieved	European 1M dataset
	Precision		Recall		F-measure
	RVM	NSVM-CNN	RVM	NSVM-CNN	RVM	NSVM-CNN
10	99.5	99.8	62.34	66.21	95.78638	79.60675
20	99.1	99.25	65.02	69.67	78.52159	81.87009
30	98.65	98.98	67.84	73.23	80.39421	84.17984
40	98.23	98.54	70.56	76.72	82.12701	86.2717
50	97.8	98.13	73.31	80.23	83.80244	88.28179
60	97.38	97.72	76.06	83.74	85.40963	90.19148
70	96.95	97.32	78.81	87.25	86.9439	92.01029
80	96.53	96.91	81.56	90.76	88.41582	93.73423
90	96.1	96.51	90	84.31	94.27	95.37685
100	95.91	96.023	100	95.91	97.73	98.35769

Back to article page