Table 2.

Summary of the application of deep learning algorithms in breast cancer histopathology for detection.

Model	Strategy	Advantages	Publication
RNN	Development of decision support systems for pathology	RNN allows neurons in the hidden layer to communicate with each other, storing the previous output as information in the hidden layer	[55]
	Propose a SmallMitosis framework for the detection of mitotic cells from hematoxylin and eosin (H&E)-stained breast histological images		[56]
Inception	Histologic identification of tumor cells in lymph nodes	Inception increases the width of the network by pooling each layer with a different convolution to extract features from the previous layer, and by adding a 1*1 convolution after the pooling layer before the 3*3 and 5*5 convolutions, which effectively avoids complex parameters and computational effort	[57]
	Improve the computer-aided diagnosis method based on deep learning		[58]
ResNet	Detection of invasive ductal carcinoma in breast histological images and the classification of lymphoma subtypes	The main feature of ResNet is the residual block, the purpose of the residual block is to preserve the characteristics of the parameters before the current layer is trained and to pass these parameters into the subsequent layers together with the trained data	[59]
	Diagnostic breast cancer whole-slide tissue images		[60]
	Propose an automatic detection method for invasive ductal carcinoma (IDC) based on deep transfer learning technology		[61]
	Propose Mask RCNN, a multi-task deep learning framework for object detection and instance segmentation, to automatically detect mitosis		[62]
DCNN	Propose an accurate method for detecting the mitotic cells from histopathological slides using a multi-stage deep learning framework		[63]
	Present an SSAE for efficient nuclei detection on high-resolution histopathological images of breast cancer		[64]
	Introduce deep learning as a technique to improve the objectivity and efficiency of histopathologic slide analysis		[65,66,67,68,69,70]
Semi-Supervised Learning	Present a semi-supervised deep learning strategy for breast cancer diagnosis	Semi-supervised learning is to use a large number of unlabeled samples and a small number of labeled samples to train the classifier, solving the problem of insufficient labeled samples	[71,72]
YOLO	A fast lesion detection method based on yolo is proposed	Simple structure and fast speed	[50]
Faster RCNN	A fast detection method of breast tumor based on Faster RCNN is proposed	Faster RCNN realizes object detection performance with high accuracy through second-order network and Region Proposal Network	[52]
Single Shot multibox Detector (SSD)	An automatic detection method of breast cancer lesion based on SSD is proposed	One stage, good at detecting small objects	[53]