Table 1.

Different classification techniques for brain tumor diagnosis

Reference	Method	Number of images in the dataset	Limitations	Accuracy %
Cheng et al. [7]	BoW, intensity histogram and GLCM	3064	Elevated computational complicatedness	91.28
	Ring from partition for classification
Ismael et al. [8]	The histogram and the GLCM for feature extraction	3064	Elevated computational complicatedness	91.9
	ANN for classification
Ari et al. [9]	ELM-LRF for classification	108	Small dataset	97.18
			Inappropriate for another training dataset
	Watershed segmentation for segmentation
Gumaei et al. [10]	(PCA) with GIST descriptors for feature extraction	3064	Elevated computational complicatedness	94.23
	Regularized extreme learning machine for classification
Sajjad et al. [11]	VGG19 with data augmentation	3064	Elevated computational cost	94.58
			Large storage required
Kutlu et al. [12]	Based on AlexNet	300	Small dataset	98.6
			Elevated computational cost
			Large storage requirements
Swati et al. [13]	VGG with fine tuning	3064	Elevated computational cost	94.82
			Large storage requirements
Ghosal et al. [14]	Based on AlexNet	3049	Elevated computational cost	93.83
			Large storage requirements
Anaraki et al. [15]	CNN with Genetic Algorithm	3064	Elevated computational cost	94.2
			Large storage requirements
Deepak et al. [16]	GoogleNet with Transfer Learning	3064	Time-consuming	97.1
			Elevated computational cost
			Large storage requirements
Alshayeji et al. [17]	Aggregation of two paths from CNN	3064	Time-consuming	97.37
			Elevated computational cost
			Large storage requirements
Kakarla et al. [18]	Average pooling convolutional neural network	3064	Time-consuming	97.42
			Elevated computational cost
			Large storage requirements
Kumar et al. [19]	ResNet-50 with Global Average Pooling at the output layer	3064	Time-consuming	97.48
			Elevated computational cost
			Large storage requirements