A State-of-the-Art Review for Gastric Histopathology Image Analysis Approaches and Future Development

Table 2

Classification-based gastric cancer image analysis.

Aim	Year	Reference	Team	Dataset	Method	Evaluation
Machine learning	2015	[27]	Sharma et al.	5,541 30x and 3,730 40x images	Adaboost
	2017	[32]	Sharma et al.	795 images	Random forest
	2018	[47]	Liu et al.	560 cancer and 140 noncancer	CNN and SVM
Deep learning	2017	[22]	Garcia et al.	3,275 images	DCNN
	2017	[23]	Sharma et al.	21,000 images	Proposed CNN	Accuracy of cancer classification is 69.9%; accuracy of necrosis detection is 81.4%
	2018	[24]	Li et al.	560 cancer and 140 noncancer	Proposed CNN	Accuracy of patch level classification is 97.93%; accuracy of slice level classification is 100%
	2018	[48]	Liu et al.	1.2 million images	ResNet
	2019	[51]	Wang et al.	608 whole slide images	ResNet
	2020	[52]	Song et al.	2,123 digital slides	DeepLab-V3 and ResNet	, ,