Computational Intelligence Method for Detection of White Blood Cells Using Hybrid of Convolutional Deep Learning and SIFT

Table 4

Comparing the accuracy of the proposed method in detecting the white blood cells with four baseline methods.

Reference	Segmentation method	Classification method	Sample size	Accuracy
The proposed method	Gram-Schmidt orthogonalization	WTPSSR	260	97.14%
The proposed model by Rezatofighi et al. [17]	Gram-Schmidt orthogonalization and snake	SVM	400	86.10%
The proposed model by Zhang et al. [26]	Histogram threshold	Distance classifier	199	92.46%
The proposed model by Balki et al. [6]	Entropy threshold and iterative threshold	Distance classifier	71	90.14%
The proposed model by Horne et al. [2]	Gram-Schmidt orthogonalization and snake	LVQ	400	94.10%