Method and device for segmenting key anatomical structure of image under assistance of global visual angle

The invention relates to an image key anatomical structure segmentation method under the assistance of a global view angle, computer equipment and a storage medium. The method comprises the following steps: inputting a global view angle single-frame image into a convolutional neural network to obtain a mask of a key anatomical structure; inputting the global view angle single-frame image into a ResNet18 framework to obtain a single-frame image feature map, and processing the single-frame image feature map by adopting a mask to obtain a mask-processed feature map; pooling the feature map subjected to mask processing to obtain a global context feature map; inputting the local view angle single-frame image into a ResNet18 framework to obtain a local context feature map; inputting the global context feature map and the local context feature map into a classification network to obtain a feature map of classification information of the key anatomical structure, and performing cross attention fusion on the feature m