Three-dimensional virtual reconstruction method and system for medical image

Abstract

The embodiment of the invention provides a three-dimensional virtual reconstruction method and system for a medical image. The method comprises the following steps: carrying out deep learning training on multiple groups of original medical images to obtain multiple groups of region feature data of regions of interest; carrying out gradient training learning according to the multiple groups of region feature data to obtain multiple groups of two-dimensional convolution kernel models; forming multiple groups of three-dimensional convolution kernel models by adopting the multiple groups of two-dimensional convolution kernel models; performing convolution processing on the multiple groups of original medical images by adopting the multiple groups of three-dimensional convolution kernel models to generate multiple groups of preprocessed medical images; and performing three-dimensional virtual reconstruction by adopting the multiple groups of preprocessed medical images. By adopting the technical scheme of the invention, the boundary of each organ in each preprocessed medical image can be clearer; and the fault data of each preprocessed medical image can realize smooth transition, so that the technical problem that the medical image has a fault during three-dimensional virtual reconstruction is solved.

Description

technical field [0001] The invention relates to the technical field of three-dimensional virtual reconstruction, in particular to a method and system for three-dimensional virtual reconstruction of medical images. Background technique [0002] The inherent characteristics of medical scanning image data, such as CT radiation damage to patients, crosstalk of MRI scanning signals, intensive scanning takes a long time, and a large amount of image data, are not suitable for intensive layer-by-layer scanning. When the scanning interval is too close, CT scanning will become The time that the patient is exposed to radioactive rays is doubled, and the order of magnitude of the scanned image data is increased at the same time, which leads to a huge consumption of processing power in the post-processing. Usually, skip scans are usually used at large intervals during scanning to reduce the radiation exposure time of patients. [0003] At present, 3D virtual image reconstruction based on...

AI Technical Summary

Problems solved by technology

Due to the problem of medical scanning interval, it will lead to the problem of jumping between layers during the stacked 3D reconstruction of virtual images.

Between different scanning image layers, the image data also has the same data transition problem as in the scanning layer

Images