Skip to main content

Table 1 Selected feature subsets from the original features using PCA method. Only 33 features were selected as the best feature set for DC characterization and these were used as the input for classification model (FOS: first order statistics, FD: fractal dimension, GLCM: gray level co-occurrence matrix, GLRLM: gray level run length matrix, and LTE: Law’s texture energy)

From: Automated classification of dense calcium tissues in gray-scale intravascular ultrasound images using a deep belief network

No.

Feature Group

Feature Subsets

1

FOS

Mean

2

Variance

3

Intensity

Intensity

4

FD

Fractal Dimension

5

GLCM

Difference Variance

6

Contrast

7

Sum Variance

8

Autocorrelation

9

Cluster Prominence

10

Sum of Squares

11

Sum Average

12

Entropy

13

Energy

14

Homogeneity

15

Maximum Probability

16

Sum Entropy

17

Dissimilarity

18

Difference Entropy

19

GLRLM

SRE

20

LRE

21

GLN

22

HGRE

23

LTE

(SSV) R5S5/S5R5

24

(SSV) E5E5

25

(SSV) E5S5/S5E5

26

(SSV) E5R5/R5E5

27

(SSV) S5S5

28

(SSV) R5R5

29

(SAV) R5S5/S5R5

30

(SAV) S5S5

31

(SAV) R5R5

32

(SAV) E5R5/R5E5

33

(SAV) L5S5/S5 L5