End-to-end image compression method based on context clustering transformation

The invention particularly relates to an end-to-end image compression method based on context clustering transformation. The method comprises the following steps: inputting an image into a trained image compression model, and outputting a reconstructed image; the model training process comprises the following steps: encoding an image: converting the image into feature points containing coordinates, then dividing the feature points into a plurality of clusters according to similarity, and finally learning image features in a manner of performing multiple aggregation and redistribution on the feature points in each cluster to generate potential representation; the entropy model completes probability modeling of each potential representation to be coded; entropy coding and entropy decoding are carried out on the potential representation to be coded through an entropy coder, and an entropy value feature map is generated; carrying out decoding reconstruction on the entropy feature map through a main decoder to gen