VVC intra-frame coding fast block partitioning method based on graph neural network

The invention provides a VVC intra-frame coding fast block partitioning method based on a graph neural network, which comprises the steps of building a coding unit partitioning structure prediction network model based on the graph neural network and provided with a texture feature encoder, a multi-way tree feature encoder and a label predictor, and setting a fast block partitioning decision method based on multiple thresholds. Training and parameter optimization of a coding unit division structure prediction network model, and a VVC intra-frame coding fast block division method flow based on a graph neural network. According to the method, the convolutional neural network and the graph neural network are used for learning important texture information of the coding unit, a cross attention and multi-way tree feature encoder is introduced to realize interactive fusion of multi-dimensional features, quantization parameters of the coding unit are normalized and then input into the network model, accurate predicti