THREE-DIMENSIONAL RECONSTRUCTION METHOD AND SYSTEM, AND STORAGE MEDIUM

The content of the present disclosure relates to a three-dimensional reconstruction method and system, and a storage medium. Various embodiments regarding three-dimensional reconstruction are described. In an embodiment, a method for training a three-dimensional reconstruction model comprises: on the basis of an image that is obtained by photographing a target object at a plurality of angles of view, generating an initial voxel envelope of the target object; performing random sampling on points in the initial voxel envelope, so as to obtain a set of sampling points; performing global feature extraction on the image, so as to obtain a global feature map; on the basis of a geometric association, determining, from the global feature map, a global feature corresponding to the sampling points; performing coding processing on geometric information related to the sampling points, so as to generate geometric coded information; and training a model at least on the basis of the global feature and the geometric coded in