An automatic steering method from spect 3D reconstruction image to standard view

Abstract

The invention discloses an automatic steering method from a SPECT three-dimensional reconstruction image to a standard view, and forms a mapping database of A and P by using a rigid registration algorithm to extract a rigid registration parameter P between a SPECT three-dimensional reconstruction image A and a standard SPECT image R , use the 3-layer convolution module to extract the features of the image A, and convert it into a 6-dimensional feature vector T after three times of full connection, and apply T to A through the space transformation network to form the steering result training of the network prediction to establish a SPECT 3D reconstruction image Automatic steering model. The standard view can be obtained by using the SPECT 3D reconstruction image to be steered as input and using the automatic steering model of the SPECT 3D reconstruction image for automatic steering. The invention uses the network to extract image position features to form fully automatic steering from different angle views to standard views, which reduces the complexity of manual steering operations and improves the convenience of image operations.

Description

technical field [0001] The present invention relates to the fields of medical imaging and deep learning, in particular to an automatic steering method from a SPECT left ventricular image to a standard view based on a deep learning network. Background technique [0002] SPECT cardiac imaging can detect potential lesions that have not yet caused structural changes and provide more detailed functional activity information of myocardial tissue. It is currently an important imaging method for the diagnosis, efficacy evaluation, and prognosis of coronary heart disease. At present, the most commonly used analysis methods for clinical diagnosis of cardiac nuclear medicine are based on quantitative analysis indicators such as polar map analysis and left ventricular ejection coefficient analysis. These methods need to rotate the left ventricle on the cardiac reconstruction image to obtain a standard cardiac view for quantitative analysis. The SPECT cardiac reconstruction image is bas...

AI Technical Summary

Problems solved by technology

Clinically, doctors need to manually convert the image from the conventional view to the standard view of the heart for clinical analysis. This subjective operation is likely to introduce random errors and affect the accuracy of the analysis, and it takes a long time for manual operation.

Images