Image processing method and device, storage medium and program product

Abstract

The embodiment of the invention provides an image processing method and device, a storage medium and a program product. The method comprises the steps of obtaining a to-be-processed image containing a target object; the inputting to-be-processed image into a mask processing model, wherein the mask processing model comprises a coding network, a decoding network and an output network, extracting coding features of the to-be-processed image through the coding network, and reconstructing a foreground feature map and a background feature map through the decoding network according to the coding features, determining a foreground mask corresponding to the to-be-processed image through an output network according to the foreground feature map and the background feature map; and performing background replacement on the to-be-processed image according to the foreground mask. According to the embodiment of the invention, the coding features to be processed are accurately extracted through the mask processing model, then the foreground features and the background features are effectively reconstructed, and the accuracy of the foreground mask can be effectively improved, so that after background replacement is carried out on the image to be processed according to the foreground mask, a high-quality image can be obtained, and the aesthetic needs of users are met.

Description

technical field [0001] Embodiments of the present disclosure relate to the field of computer technology, and in particular, to an image processing method, device, storage medium, and program product. Background technique [0002] In some scenarios, it is necessary to replace the background in the image. For example, in a live broadcast scene, to replace the portrait background in the live broadcast screen with a specific background image, it is usually necessary to use a portrait matting algorithm to separate the foreground portrait from the background, and then combine it with the preset The background image is fused together to achieve background replacement. [0003] The current mainstream matting algorithm is to use semantic segmentation or other matting algorithms, input an RGB image, the image attribute shape is (3, H, W), and output a transparency map, the shape is (1, H, W). However, the accuracy of the foreground portrait obtained by the portrait matting algorithm ...

AI Technical Summary

Problems solved by technology

However, the accuracy of the foreground portrait obtained by the portrait matting algorithm in the prior art is low, and it usually cannot meet people's aesthetic requirements.

Images