IoMT-Based Automated Detection and Classification of Leukemia Using Deep Learning

Table 5

A comparison of the proposed models with the previous approaches for automated detection of leukemia and its subtypes using the same datasets with respect to average accuracy.

Reference	Dataset	Classification	Classifier	Accuracy (%)
Ahmed et al. [39]	ALL-IDB	Leukemia vs healthy	CNN	88.25
			Naive Bayes	69.69
			Decision tree	62.94
			KNN	58.57
			SVM	50.09
	ALL-IDB, ASH image bank	Leukemia subtypes classification	CNN	81.74
			Naive Bayes	52.68
			Decision tree	45.92
			KNN	43.51
			SVM	20.84
Shafique et al. [26]	ALL-IDB	Acute lymphoblastic leukemia detection	AlexNet	99.50
		Subtypes of acute lymphoblastic leukemia	AlexNet	96.06
Jothi et al. [34]	ALL-IDB	Acute lymphoblastic leukemia detection	Jaya, SVM	99.00
			Jaya, decision tree	98.00
Acharya et al. [40]	ALL-IDB	White blood cells	K-medoids algorithm	98.60
Mishra et al. [35]	ALL-IDB1	Acute lymphoblastic leukemia detection	DOST, PCA, LDA	99.66
Tuba et al. [41]	ALL-IDB2	Acute lymphoblastic leukemia detection	GAO-based methods	93.84
Al-jaboriy et al. [42]	ALL-IDB1	Acute lymphoblastic leukemia detection	GA and ANN	97.07
Jha et al. [43]	ALL-IDB2	Acute lymphoblastic leukemia detection	SCA-based deep CNN	98.70
Pansombut et al. [44]	ASH image bank, ALL-IDB1	Lymphoblast cells	CNN-based convnet	81.74
Vogado et al. [36]	Heterogeneous database ALL-IDB1, ALL-IDB2	Diagnose leukemia (pathological or not)	Pre-trained CNN with SVM	99
Thanh et al. [45]	ALL-IDB1	Diagnose leukemia (normal vs abnormal)	CNN	96.60
Moshavash et al. [46]	ALL-IDB1, ALL-IDB2, Dr. Juan Bruno Zayas Alfonso Hospital, Santiago de Cuba	Acute lymphoblastic leukemia detection	Two ensemble classifiers with SVM	89.81
Umamaheswari et al. [47]	ALL-IDB2	Acute lymphoblastic leukemia	Customized KNN	96.25
Agaian et al. [48]	ALL-IDB1	Acute lymphoblastic leukemia	Cell energy feature with SVM	94.00
Rawat et al. [49]	ASH image bank	ALL	GA with SVM	97.10
		AML		98.50
		Healthy, AML,ALL		99.50
Proposed work	ALL-IDB, ASH image bank	Healthy, ALL, AML, CLL and CML	ResNet-34	99.56
			DenseNet-121	99.91