Lightweight retinal vessel segmentation method based on graph convolution network and partial convolution

The invention provides a light-weight retinal vessel segmentation method based on a graph convolutional network and partial convolution, which comprises the following steps of: constructing a light-weight retinal vessel segmentation network model based on the graph convolutional network and the partial convolution, wherein the light-weight retinal vessel segmentation network model is provided with a feature encoder, a multi-scale feature fusion device, a feature decoder and a label predictor; training and parameter optimization of a retinal vessel segmentation network model, and automatic semantic segmentation of a retinal vessel structure. According to the method, a symmetric codec deep learning model is built, partial convolution is adopted to replace conventional convolution to reduce the calculation complexity of the model, a K-nearest neighbor algorithm is used to convert a feature graph into a graph structure, then graph convolution is used to extract global features of the image, a multi-scale feature