Skip to main content
. 2024 Feb 14;10(4):e26192. doi: 10.1016/j.heliyon.2024.e26192

Table 7.

Comparison of results with previous studies.

Author Features Used Performance
Teramoto et al. [73]
  • 1

    Shape

  • 2

    Intensity

 Sensitivity = 83.00%,
Orozco et al. [74]
  • 1

    Texture

 Sensitivity = 84.00%,
Guo et al. [75]
  • 1

    Texture

  • 2

    Shape

 Sensitivity = 94.00%,
Messay et al. [76]
  • 1

    Shape

  • 2

    Gradient

  • 3

    Intensity

 Sensitivity = 82.00%,
Retico et al. [77]
  • 1

    Texture

  • 2

    Morphology

 Sensitivity = 72.00%,
Hussain et al. [78] RICA features and SVM Accuracy = 99.77%
Dandil et al. [79]
  • 1

    Statistical

  • 2

    Shape

  • 3

    GLCM

  • 4

    Energy

 Sensitivity = 97.00%,
Specificity = 94.00%
Accuracy = 95.00%
This study Single features
  • a)

    Haralick (SVM polynomial)

  • b)

    GLCM (SVM polynomial)

  • c)

    SIFT (SVM Gaussian)

Hybrid features approach
(GLCM + autoencoder, Haralick + Autoencoder, GLCM + Haralick) features		 Single Features
  • a)

    Accuracy = 99.89%

  • b)

    Accuracy = 98.69%

  • c)

    Accuracy = 98.39%

Hybrid features approach
Specificity = 100%
Sensitivity = 100%,
AUC = 1.00
Accuracy = 100%