Color Image Retrieval Method Using Low Dimensional Salient Visual Feature Descriptors for IoT Applications

Algorithm 2

Texture information extraction.

	Require: an RGB color image
	Ensure: texture feature descriptor
(1)	Consider V-component of the preprocessed HSV color image and apply the Sobel operator to obtain magnitude of gradients in x-and y-directions.
(2)	Divide the obtained magnitude of gradients into nonoverlapping blocks of fixed size .
(3)	Consider FV texture as an empty row vector and .
(4)	Collect DC coefficients and select some significant AC coefficients from all blocks. Correspondingly, determine the relation between two adjacent blocks using DC and AC coefficients throughout the image using equation 6 and 8, where two row vectors have been achieved.
(5)	Arrange the DC and AC row vectors into the form of matrices and apply the GLCCM approach on them. Then, GLCCM have been constructed for different directions and statistical parameters using equation 10 have been computed and collective integration of them represents the texture feature descriptor FV texture of the image plane.