Image recognition method and device, electronic equipment and readable storage medium

Abstract

The invention discloses an image recognition method and device, electronic equipment and a readable storage medium. The image recognition method comprises the steps of determining a first feature mapof a current frame image through a convolutional neural network based on the type of the current frame image; determining a second feature map of a key frame image before the current frame image; performing feature alignment on the first feature map and the second feature map to obtain a first alignment feature map; fusing the first feature map and the first alignment feature map to obtain a firstfused feature map; and based on the first fusion feature map, identifying the content in the current frame image. Thus, the features in the current frame of image can be fused on the basis of featurealignment for image recognition, the number of the features for image recognition is increased, and the accuracy of image recognition can be improved by performing image recognition on the basis of the increased number of the features.

Description

technical field [0001] The present application relates to the field of image processing, in particular to an image recognition method, device, electronic equipment and readable storage medium. Background technique [0002] The progress of the existing image detection and segmentation technology is based on the image level. How to efficiently solve the video-based detection and segmentation is still a challenging problem due to two reasons. First, the image-based detection and segmentation model generally has a large amount of calculation and parameters, and directly applying this model to the video frame by frame will generate a huge amount of calculation; second, the objects in the video will have some motion blur , out of focus and rare pose issues. [0003] In order to solve the above problems, the existing technology usually selects some sparse key frames in the video, and the remaining other frames are regarded as non-key frames, and then only extract features on these...

AI Technical Summary

Problems solved by technology

[0004] However, the optical flow network in the prior art is used to capture the corresponding points of pixels between two pictures and then output the corresponding offset. Directly using the optical flow network on the feature level with more abstract semantic information will It is difficult to find the spatial correspondence and offset of two high-level features, resulting in a decrease in the accuracy of image recognition

Images