Histopathological Tissue Segmentation of Lung Cancer with Bilinear CNN and Soft Attention

Table 3

Results on colorectal cancer dataset (all models were pretrained on ImageNet).

Model	Average precision	Average recall	Average F1
VGG19 [30]	0.9650	0.9660	0.9655
DeepTissue Net [14]	0.9770	0.9775	0.9772
EfficientNet [31]	0.9779	0.9784	0.9781
ResNet50 [29]	0.9736	0.9746	0.9741
ResNet50+bilinear pooling module	0.9764	0.9771	0.9767
ResNet50+attention module	0.9789	0.9794	0.9791
ResNet50+bilinear pooling module+attention module	0.9823	0.9826	0.9824