Emerging Feature Extraction Techniques for Machine Learning-Based Classification of Carotid Artery Ultrasound Images

Table 2

List of selected significant features.

Sl. no	Selected features	Description
1	Texture	Spatial arrangement of image intensity
2	Spatial structure	Exploit location information
3	Skewness	Extent to which a distribution differs from a normal distribution
4	Kurtosis	Pixel intensity distribution
5	Histogram	Pixel distribution as a function of tonal variation
6	Correlogram	Spatial correlation of intensity changes with distance
7	HOG	Count incidences of gradient alignment in localized regions of an image
8	Gabor wavelet	Frequency-wise intensity variation check in specific direction
9	Angular 2nd moment	Textural uniformity in image
10	Shape	Shape characteristics
11	Sharpness	Degree of clarity in both coarse and fine image detail
12	Length irregularity	Irregularities of the length of structures in an image
13	Mean probability density function	Probability that the region brightness is less than or equal to a specified brightness value
14	Grayscale median	Median of grayscale intensities
15	Multiregion histogram	Checks whether plaque outer region signifies disease progression
16	Arterial wall ROI’s randomness	Randomness present in the artery wall
17	Absolute gradient	Directional change in intensity
18	Angular and radial sum of discrete Fourier transform for Fourier power spectrum	Fourier power spectrum’s Fourier transform
19	Coarseness	Type of texture feature
20	Convexity	Convex curves present in an image
21	Connectivity	Connectivity among pixels
22	Plaque volume	Plaque volume measure