Involvement of Machine Learning for Breast Cancer Image Classification: A Survey

Table 2

Feature descriptor.

Feature category	Feature description
Texture	Haralick texture features [7]
	Angular Second Moment (ASM), Contrast, correlation, Sum of Squares of Variances (SSoV), Inverse of Difference (IoD), Sum of Average (SoA), Sum of Variances (SoV), Sum of Entropy (SoE), Entropy, Difference of Variance (DoV), Difference of Entropy (DoE), Gray-Level Concurrence Matrix (GLCM).
	Tamura features [8]
	Coarseness, Contrast, directionality, line-likeness, roughness, regularity.
	Global texture descriptor
	Fractal dimension (FD), Coarseness, Entropy, Spatial Gray-Level Statistics (SGLS), Circular Moran Autocorrelation Function (CMAF).
Detector	Single scale detector
	Moravec’s Detector (MD) [9], Harris Detector (HD) [10], Smallest Univalue Segment Assimilating Nucleus (SUSAN) [11], Features from Accelerated Segment Test (FAST) [12, 13], Hessian Blob Detector (HBD) [14, 15].
	Multiscale detector [8]
	Laplacian of Gaussian (LoG) [9, 16], Difference of Gaussian (DoG) Contrast [17] Harris Laplace (HL), Hessian Laplace (HeL), Gabor-Wavelet Detector (GWD) [18].
Strutural	Area, bounding box, centroid, Convex Hull (CH), eccentricity, Convex Image (CI), compactness, Aspect Ratio (AR), moments, extent, extrema, Major Axis Length (MaAL), Minor Axis Length (MiAL), Maximum Intensity (MaI), Minimum Intensity (MiI), Mean Intensity (MI), orientation, solidity.