. 2023 Nov 3;9(11):e21703. doi: 10.1016/j.heliyon.2023.e21703

Table 5.

Five sets of extracted geometrical features from ROI images.

Category	Feature name	Feature descriptions
Set1: Morphological features	Kurtosis	The density distribution of pixels
	Skewness	The degree of symmetry in the image's pixel distribution.
	Extent	The proportion of the total area of the ROI to the total area of the convex hull.
	Solidity	Using the pixels that make up the Convex Hull to contrast object regions with its Convex Hull.
	Circularity	The measurement of the ROI's roundness
	Major axis length	Calculate the longest length of the ROI object
	Minor axis length	Calculate the minimum length of the ROI object
	Equivalent diameter	This is the radius of a circle with the same circumference as the ROI region.
Set 2: GLCM features	Energy	The square root of an angular second moment is used to calculate energy. When the window is neatly arranged, energy has a larger value.
	Contrast	Contrast is a unique GLCM moment that is used to quantify the spatial frequency of an image. It's calculated by taking the range from the highest and lowest neighboring pixel values.
	Dissimilarity	Dissimilarity is a linear way to measure the differences between parts of an image.
	Homogeneity	It assesses image homogeneity by assuming bigger values for smaller variances in gray tone within-pair components. Homogeneity in the GLCM is particularly sensitive to the presence of near diagonal components.
	correlation	It is a measure of the linear relationship between the gray tones of the image.
	Entropy	It evaluates the randomness of intensity levels in the neighborhood.
Set 3: Statistical features	Mean	The sum of all pixels divided by the total number of pixels
Set 3: Statistical features	Standard deviation	The measurement of dispersion in the image's gray intensity level.
Set 4: Texture features	Texture Energy	It shows how rough the surface is in the defect image.
Set 4: Texture features	Texture Entropy	This denotes the textural complexity of the fault image
Set 5: LBP features	LBP Energy	The LBP features are produced by contrasting the central pixel with its surroundings in a limited area of the image. These features define the image local texture properties and provide important advantages, including rotation and gray invariance.
Set 5: LBP features	LBP Entropy