| | Studies | Dataset | Feature extraction method | Number of features | Classifier | Tumor types | Performance |
| | Sarkar et al. [8] | Local dataset:
60 MRI T2 images | Genetic algorithm | 13 features | SVM | Benign,
malignant | SPE 100%, SEN 98%, ACC 98.30% |
| | Hamid et al. [9] | Local dataset:
60 MRI T2- FLAIR images | GLCM | 5 features | SVM | Benign,
malignant | ACC 95% |
| | Ansari et al. [10] | Local dataset:
200 MRI | GLCM and DWT | 12 features | SVM | Benign,
malignant | ACC 98.91% |
| | Li et al. [11] | Local dataset:
135 MRI T1 and T2 images | Gabor transform, texture, and DWT | 80 best-ranked features | SVM | Benign,
malignant | SPE 80%, SEN 93%, ACC 88% |
| | Alves et al. [12] | Local dataset:
67 patients
T1, T1C+, T2, DWI, T2-FLAIR | Genetic algorithm, GLCM, GLRL and DWT | Five best-ranked features | SVM
kNN
RF | Brain tumors and inflammatory lesions | SPE 83.70%, SEN 91.20%, ACC 82.70% |
| | Kang et al. [13] | Local dataset: 6,517 MRI | CNN | Top-3 deep features | SVM
kNN
NB
RF | Benign,
malignant | SVM: ACC 98.50%
kNN: ACC 98.50%
NB: ACC 90.20%
RF: ACC 97.17% |
| | Jena et al. [14] | BraTS 2017 and 2019 | Genetic algorithm, GLCM, GLRL and DWT | 471 features | SVM
kNN
RF | LGG, HGG | SVM: ACC 94.25%
kNN: ACC 87.88%
RF: ACC 97% |
| | Nanmaran et al. [15] | Local dataset: 200 MRI | Discrete cosine transform | 6 features | SVM
kNN | Benign,
malignant | SVM: ACC 96.8%
kNN: ACC 91.75% |
| | Susanto et al. [16] | BraTS 2019 | GLCM and DWT | 16 features | SVM | LGG, HGG | SVM: ACC 98.65% |
| | Aamir et al. [17] | Local dataset: 3,064 MRI | Multiple deep neural networks | Deep features | SVM | Benign,
malignant, pituitary | SVM: ACC 98.98% |
|
|
