Training sample generation method, image processing method, device, equipment and medium

Abstract

Embodiments of the present disclosure disclose a training sample generation method, an image processing method, a device, an electronic device, and a medium. A specific implementation of the above training sample generation method includes: acquiring a face image with a mouth displayed and labeling information of the face image; extracting the mouth image from the face image; transforming the labeling information to obtain the labeling information after transformation; Sorting the key points of each tooth in the at least one tooth with the transformed annotation information to obtain a ranking number; generating training samples according to the mouth image, the sorting number, and the transformed coordinate information. This embodiment provides a training sample generation method. The training sample generated by the method can conveniently locate the position of the key point of a single tooth, improve the accuracy of the subsequent prediction of the key point information of the tooth, and prepare for the subsequent prediction of a single tooth.

Description

Training sample generation method, image processing method, apparatus, equipment and medium technical field Embodiments of the present disclosure relate to the field of computer technology, in particular to a training sample generation method, an image processing method Methods, apparatus, electronic devices, and computer-readable media. Background technique [0002] At present, various short video platforms add special effects to scenes or characters in the video to add interest to the short video. Adding special effects to teeth (for example, tiger teeth effects, rabbit teeth effects, and tooth decay effects) requires processing a single tooth. With artificial With the development of intelligent technology, network models can be used to extract the information needed in a single tooth. For different training effects, the The network model used needs to generate different training samples. Generally speaking, training samples can be obtained by direct shooting or ...

AI Technical Summary

Problems solved by technology

Generally speaking, the way to obtain training samples is direct shooting or database. Such acquisition methods can no longer meet the needs of training network models.

Images