Image segmentation method, device, equipment and storage medium

Abstract

The invention discloses an image segmentation method, device, electronic equipment and storage medium. The method includes: precoding the image data of the image to be processed to obtain a precoded feature map; performing a first decoding operation on the precoded feature map to obtain geometric edge information; performing feature coding on the precoded feature map to obtain The first feature map; performing context semantic capture on the first feature map based on several separable atrous convolutions to obtain semantic information of the first feature map; fusing the geometric edge information with the context semantic information; The fusion result is decoded to obtain a first mask of the image to be processed, and the first mask is used to perform image segmentation on the image to be processed. The technical scheme of the invention can achieve the purpose of reducing the image segmentation calculation amount and cache bandwidth of the device, thereby improving the image segmentation speed and image segmentation accuracy.

Description

technical field [0001] The embodiments of the present application relate to the technical field of information processing, and in particular, to an image segmentation method, device, electronic equipment, and storage medium. Background technique [0002] With the rapid development of short video content, in the scene of mobile audio and video APP or hardware image processing, the demand for image segmentation in video is increasing. [0003] However, in the image segmentation process, existing segmentation methods, such as traditional PC-side algorithms or image segmentation algorithms that rely on GPU operations, are usually based on the functions provided by some software on the mobile terminal to process the image to be processed. Since the video needs to be processed in real time during the user's use, the timeliness of the presentation and the user's perception are better. However, due to the constraints of mobile devices, such as insufficient computing efficiency, slo...

AI Technical Summary

Problems solved by technology

However, due to the constraints of mobile devices, such as insufficient computing efficiency, slow cache speed, and small bandwidth, the existing image segmentation methods can only run normally on the server side, but cannot run smoothly on the mobile side. conduct

At the same time, since the existing image segmentation methods rely on GPU operations, it will also lead to problems such as slow segmentation speed and low precision.

Images