Polyp segmentation method based on significance map guidance and uncertainty semantic enhancement

The invention discloses a polyp segmentation method based on significance map guidance and uncertainty semantic enhancement. According to a medical image to be segmented, firstly, a feature map and a saliency map are extracted through the trunk coding sub-network, higher-order feature representation is learned for the feature map through the second-order pooling convolution attention sub-network so as to enhance the nonlinear modeling capability, the feature map is guided through the uncertainty semantic enhancement sub-network by using the saliency map, and the non-linear modeling capability is enhanced. Namely, the network is guided to pay attention to learning of target area features, and finally, the significance map is subjected to up-sampling and activation functions to obtain a final prediction segmentation result of the medical image. The method tries to guide the network to pay attention to the learning of the features of the target region through the calculation of the significance and uncertainty o