Method and device for reconstructing complete three-dimensional teeth from single monocular image and storage medium

The invention relates to the field of image processing, in particular to a method and equipment for reconstructing a complete three-dimensional tooth from a single monocular image and a storage medium, which are based on a three-dimensional tooth reconstruction technology of deep learning, and are used for training by using tooth model data obtained by a large amount of scanning and segmented tooth pictures. The segmentation of the tooth picture and the reconstruction of the corresponding three-dimensional tooth can be completed at the same time, and the generation of the tooth root is completed in the post-processing process. After network training is completed, a single picture serves as input, the network firstly performs feature extraction and segmentation on the picture, then implicit reconstruction is performed on a tooth crown part model, and after a three-dimensional tooth crown model is obtained, a series of post-processing algorithms are used for generating a tooth root, so that reconstruction of a