Image labeling method and device based on multi-model fusion, computer equipment and storage medium

Abstract

The invention relates to the field of image detection, and the labeling effect is improved by fusing a binary classification result of a classification model based on a DenseNet network and a binary segmentation result of a segmentation model based on a Vnet network and an FPN network. The invention particularly discloses an image labeling method and device based on multi-model fusion, computer equipment and a storage medium. The method comprises the steps of acquiring a to-be-labeled image, and the to-be-labeled image is preprocessed to obtain a plurality of instance images; inputting each instance image into a classification model based on a DenseNet network for binary classification; splicing the binary classification results corresponding to the plurality of instance images to obtain classification result images; inputting each instance image into a segmentation model based on a Vnet network and an FPN network to carry out binarization segmentation; splicing the binarization segmentation results corresponding to the plurality of instance images to obtain a segmentation result image; calculating a binarized fusion image according to the classification result image and the segmentation result image; and extracting the contour of the fused image to mark the region of interest in the to-be-marked image according to the contour.

Description

technical field [0001] The present application relates to the technical field of image processing, in particular to an image tagging method, device, computer equipment and storage medium based on multi-model fusion. Background technique [0002] In some application scenarios, it is necessary to segment the target region of interest in the image; however, due to the fine-grained changes in some images, such as skin lesion images, it is very difficult for the system to automatically classify each region of the image. Although the existing deep convolutional neural network (CNN) is often used to classify various fine-grained objects and performs well in multiple tasks, it still suffers from low accuracy. Contents of the invention [0003] Embodiments of the present application provide an image tagging method, device, computer equipment, and storage medium based on multi-model fusion, which can better realize the tagging of the region of interest in the image to be tagged, and...

AI Technical Summary

Problems solved by technology

[0002] In some application scenarios, it is necessary to segment the target region of interest in the image; however, due to the fine-grained changes in some images, such as skin lesion images, it is very difficult for the system to automatically classify each region of the image

Although the existing deep convolutional neural network (CNN) is often used to classify various fine-grained objects and performs well in multiple tasks, there is still a problem of low accuracy

Images