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

Table 1

Results on lung cancer dataset (all models pretrained on colorectal cancer dataset).

Model	Average precision	Average recall	Average F1
ResNet50 [29]	0.9239	0.9279	0.9259
VGG19 [30]	0.9185	0.9238	0.9211
EfficientNet [31]	0.9184	0.9295	0.9239
DeepTissue Net [14]	0.9218	0.9250	0.9234
ResNet50+bilinear pooling module	0.9253	0.9286	0.9269
ResNet50+attention module	0.9268	0.9291	0.9279
ResNet50+bilinear pooling module+attention module	0.9394	0.9415	0.9404