Multi-scale and multi-feature electrocardiogram medical image fusion and classification method

Abstract

The invention claims a multi-scale and multi-feature electrocardiogram medical image fusion and classification method. The method comprises the following steps: 1, acquiring all the electrocardiogrammedical images in a database, and performing a preprocessing step including removing baselines and high frequency noise in the electrocardiogram medical images; 2, performing smoothing processing andperforming multi-scale decomposition in a multi-scale decomposition process to obtain a smooth image and a detail image; and using information entropy to perform weighted fusion to obtain the fused smooth image for the smooth image S, and using multiple features to perform fusion to obtain the fused detail image FS for the detail image; 3, performing ADMM based reconstructing on the fused smooth image FD and the detail image FS by using the simple operation at the pixel level to obtain the fused image F; and outputting the final fusion image F; and 4, adopting an Adaboos classifier to classifythe fused image F. The invention can lower the complexity, improve the quality of the fused image and improve the classification accuracy.

Description

technical field [0001] The invention belongs to the technical field of medical image processing, in particular to a multi-scale and multi-feature electrocardiogram medical image fusion and classification method. Background technique [0002] The existing electrocardiogram medical image fusion technology is mainly for two different modal medical images. According to the different medical image modalities, the medical image fusion system can be divided into three types: anatomical medical image and anatomical medical image fusion, anatomical medical image and anatomical medical image fusion. Functional medical image fusion and functional medical image and functional medical image fusion. The MRI-PET and MRI-SPECT medical image fusion system belongs to the fusion of anatomical medical images and functional medical images. The input images of this system are grayscale and pseudo-color. The MRI-PET combination kit launched by Philips combines a commercial MRI imaging scanner wit...

AI Technical Summary

Problems solved by technology

In terms of image decomposition and reconstruction methods, image fusion methods based on the frequency domain use Fourier transform and inverse Fourier transform to decompose and reconstruct image signals at different scales, but this type of method has high time complexity and runs The characteristics of long time, which contradicts the high real-time medical aided diagnosis, and has very high requirements on the hardware and software facilities of the experimental platform.

Later, researchers proposed to use spatial domain filters to process images to perform multi-scale decomposition and reconstruction of input images. Although this type of method can quickly perform image decomposition and reconstruction, the image fusion method based on spatial domain has poor noise immunity.

However, the existing methods use the same feature for fusion of medical images of different modalities.

Images