IRR-Net: A Joint Learning Framework for Image Reconstruction and Recognition of Photoacoustic Tomography

Table 2

Two-step methods for comparison.

Name of method for comparison		Image reconstruction method	Classifier
HH	5 DGD + GoogLeNet	Five cascaded DGD units	GoogLeNet
	DGD + U-Net+ResNet	A single DGD unit followed by a U-Net	ResNet50
LH	1 DGD + GoogLeNet	A single DGD unit	GoogLeNet
	TR + GoogLeNet	TR	GoogLeNet
	TR + AlexNet	TR	AlexNet
	TR + ResNet	TR	ResNet
HL (5 DGD + SVM)		Five cascaded DGD units	SVM
LL (TR + SVM)		TR	SVM

Note: HH, LH, HL and LL refer to “High-quality reconstruction + High-performance classifier”, “Low-quality reconstruction + High-performance classifier”, “High-quality reconstruction + Low-performance classifier”, and “Low-quality reconstruction + Low-performance classifier,” respectively.