Noise Resilient Local Gradient Orientation for Content-Based Image Retrieval

Table 11

Performance of NRLGO is compared with uniform pattern-based state-of-the-art methods on Corel dataset with 5000 images and Corel dataset with 10000 images.

Dataset	Performance	CPV-THF [42]	GLCM [50]<	SQs [49]	GMMs [49]	SED [46]	CMSD [12]	Proposed
Corel 5k	Precision (%)	63.90	38.0	41.67	51.80	56.3	63.14	65.8
	Recall (%)	7.66	1.08	5.13	6.22	21.3	7.57	7.89
Corel 10k	Precision (%)	52.28	31.9	37.64	47.25	46.4	50.25	53.07
	Recall (%)	6.27	.63	4.51	5.67	18.3	6.03	6.36