Image processing method and device, electronic equipment and storage medium

Abstract

The invention relates to an image processing method and device, electronic equipment and a storage medium, and the method comprises the steps: carrying out the feature extraction of a color image anda corresponding depth image, obtaining N levels of first feature maps of the color image and N levels of second feature maps of the depth image, wherein the scales of all levels of first feature mapsin the N levels of first feature maps are different, the scales of all levels of second feature maps in the N levels of second feature maps are different, and N is an integer greater than or equal to2; fusing the nth-level first feature map of the N levels of first feature maps with the nth-level second feature map of the N levels of second feature maps to obtain an nth-level third feature map ofthe N levels of stage third feature maps, n being an integer and being greater than or equal to 1 and less than or equal to N; performing fusion processing on the N levels of third feature maps, andobtaining a fusion feature map according to a fusion processing result; and determining a processing result of the color image and the corresponding depth image according to the fusion feature map. According to the embodiment of the invention, the accuracy of image processing can be improved.

Description

Technical field[0001]The present disclosure relates to the field of computer technology, and more particularly to an image processing method and device, an electronic device, and a storage medium.Background technique[0002]In the field of computer visual, the image is typically required, such as a feature map of the image, and analyzes the object information in the image scene according to the feature map, and the processing result of the image is obtained.[0003]The use of color images and corresponding depth images (eg, RGB-D images) are semantically splitted and instance segmentation, and the characteristic map of deeper network extraction is employed. As the semantic hierarchy is raised, the resolution is lowered, Thereby, there may be too much loss of image details, affecting the accuracy of the image after division.Inventive content[0004]This disclosure proposes an image processing technology solution.[0005]According to an aspect of the present disclosure, an image processing me...

AI Technical Summary

Problems solved by technology

[0003] In the schemes of semantic segmentation and instance segmentation using color images and corresponding depth images (such as RGB-D images) in related technologies, feature maps extracted by deeper networks are used. As the semantic level increases, the resolution decreases. As a result, too much image detail may be lost, affecting the accuracy of the segmented image

Images