Image segmentation method based on ResNet and UNet models

Abstract

The invention discloses an image segmentation method based on ResNet and UNet models, and the method comprises the steps of adjusting the size of an original RGB three-channel image, and correspondingly adjusting the size of a label image; taking the adjusted RGB image as the input of a UNet image segmentation module; taking the adjusted RGB image as the input of a ResNet feature extraction module, and reserving the output results of the first three layers to replace the output results of the third, fourth and fifth layers of UNet; obtaining an image segmentation training model based on ResNetand UNet, and training the model; and taking model parameters obtained by training as a prediction model, and carrying out image segmentation. According to the method, the quality of image segmentation is improved by utilizing the advantages of ResNet at the aspect of feature extraction, and the problems of inaccurate feature extraction, poor region consistency and fuzzy boundary which are easilygenerated when a single UNet model is applied to image segmentation are solved.

Description

technical field [0001] The invention belongs to the technical field of image processing, in particular to an image segmentation method based on ResNet and UNet models. Background technique [0002] Image segmentation is the technology and process of dividing an image into several specific regions with unique properties and proposing objects of interest. Image segmentation is an important step from image processing to image analysis, and is the basis of object classification and recognition. There are four classic image segmentation methods: [0003] Threshold-based segmentation method: The principle of threshold segmentation method is to divide image pixels into several categories by setting different feature thresholds. Firstly, the feature value is found from the grayscale or color features in the original image according to certain criteria, and the image is divided into several parts. This type of method is simple to implement and has a small amount of calculation, bu...

AI Technical Summary

Problems solved by technology

[0007] The technical problem to be solved by the present invention is to provide an image segmentation method based on ResNet and UNet models for the deficiencies in the above-mentioned prior art, utilize the advantages of ResNet in feature extraction, improve the quality of image segmentation, and solve the problem of single UNet model application Due to image segmentation, the feature extraction is not accurate enough, the regional consistency is poor, and the boundary is blurred.

Images