Multi-modal medical image fusion method based on double-residual ultra-dense network

Abstract

The invention discloses a multi-modal medical image fusion method based on a double-residual ultra-dense network, and the method comprises the steps: extracting the shallow features of a first modal medical image and a second modal medical image through the convolution of a first Conv layer and the activation of a PReLU layer in the double-residual ultra-dense network; extracting deep features through residual learning and ultra-dense connection; and sequentially performing Concat layer channel dimension splicing in the double-residual ultra-dense network on the deep features, and finally performing Conv layer convolution and PReLU layer activation to obtain a fused image of the first modal medical image and the second modal medical image. The double-residual ultra-dense block provided bycombining the residual dense block and the ultra-dense connection not only applies the dense connection between layers of the same path, but also applies the dense connection between layers crossing different paths, and performs information transmission between two paths for extracting different modal image features, so that the extracted deep features are more detailed and richer, and loss of useful information of the network intermediate layer is reduced.

Description

technical field [0001] The present invention relates to the technical field of medical image fusion, in particular to a multimodal medical image fusion method based on double residual super-dense network. Background technique [0002] Image fusion is widely used in medical imaging, remote sensing, machine vision, biometrics and military applications. The purpose of fusion is to achieve better contrast and perceptual experience. In recent years, with the increasing demand for clinical applications, research on multimodal medical image fusion has attracted much attention. The purpose of multimodal medical image fusion is to provide a better medical image to help doctors perform surgical intervention. [0003] At present, there are many modalities of medical images, such as magnetic resonance (MR) images, computed tomography (CT) images, positron emission tomography (PET) images and X-ray images, etc., and images of different modalities have their own Advantages and limitati...

AI Technical Summary

Problems solved by technology

Using residual learning or dense connection for image fusion improves the spatial resolution of the fused image, but Densenet and Resnet only use the results of the last layer of the network for feature fusion, which will result in the loss of some useful information obtained by the middle layer. Fusion Image details are not clear

Images