Image segmentation model training, image segmentation method, device and electronic equipment

Abstract

The present disclosure relates to an image segmentation model training, image segmentation method, device, and electronic equipment. The image segmentation model training method includes obtaining target category feature information and its associated scene feature information representing the category features of training samples and prediction samples; The category feature information and the associated scene feature information are spliced; the first spliced ​​feature information obtained by the splicing process is input into the initial generation network for image synthesis processing; the first composite image obtained by the synthesis process is input into the initial discriminant network for authenticity discrimination; The first synthetic image is input into the classification network of the initial image segmentation model for image segmentation, and the first image segmentation result is obtained; based on the first image discrimination result, the first image segmentation result and the target type feature information, the classification network of the initial image segmentation model is trained to obtain Target Image Segmentation Model. The image segmentation accuracy of the trained target image segmentation model can be improved by using the embodiments of the present disclosure.

Description

technical field [0001] The present disclosure relates to the technical field of artificial intelligence, and in particular to an image segmentation model training, image segmentation method, device and electronic equipment. Background technique [0002] Artificial Intelligence (AI) technology is a comprehensive subject that involves a wide range of fields, including both hardware-level technology and software-level technology. Among them, the use of artificial intelligence technology for image segmentation plays an important role in many fields such as video surveillance and public security. [0003] In related technologies, due to the high cost and difficulty of constructing training samples, the zero-shot segmentation technology scheme that automatically synthesizes image pixel features based on word vectors of unknown categories is very popular in the industry. However, since only text information participates in the training of the image segmentation model in zero-shot ...

AI Technical Summary

Problems solved by technology

[0004] The present disclosure provides an image segmentation model training, image segmentation method, device, and electronic equipment to at least solve the problems of low segmentation accuracy and error-prone in related technologies

Images