Knee joint model construction method for MRI local SAR estimation

Abstract

The invention discloses a knee joint model construction method based on tissue simplification for local SAR (Synthetic Aperture Radar) estimation. The method comprises the following steps: step 1, marking a to-be-segmented knee joint low-field magnetic resonance image as muscle, fat and bone; step 2, performing data expansion on the original low-field magnetic resonance image and the labeled imageto form a data set; step 3, respectively endowing the skeleton and the background with a first sub-network and a second sub-network, endowing the muscle and the fat with a main network, and trainingthe data input into the sub-network and the main network; step 4, combining the outputs of the sub-network and the main network to form segmentation slices; and step 5, carrying out knee joint model reconstruction after extrapolation of the segmentation slices, and carrying out electromagnetic simulation by using the extrapolated model to carry out local SAR estimation. The model reconstructed bythe method is highly similar to the real situation, and the electromagnetic simulation result is more accurate.

Description

technical field [0001] The invention relates to the technical field of medical image segmentation and deep learning, in particular to a method for constructing a knee joint model for MRI local SAR estimation. Background technique [0002] The knee joint is the largest joint tissue in the human body. Affected by human aging or inappropriate exercise, the knee joint is a very prone to injury. Magnetic resonance imaging (MRI) is one of the main methods for the diagnosis of knee joint diseases. High field MRI has the advantages of high signal-to-noise ratio and high resolution. However, during the scanning process, the local specific absorption rate (SAR) in the knee joint tissue is a key safety factor that needs to be considered. If the local SAR exceeds the standard, it will cause human body Tissues are thermally damaged, and the International Electrotechnical Commission (IEC) has corresponding regulations and requirements for the local specific absorption rate of relevant par...

AI Technical Summary

Problems solved by technology

[0006] At present, there is still a problem in practical application, that is, when the axial image is scanned, due to the limitation of coil sensitivity and imaging time, the number of layers acquired in the axial direction is limited, resulting in the length of the model being not long enough compared with the length of the coil. There will be errors in the electromagnetic simulation, which cannot effectively reflect the actual situation of the human leg placed in the coil, so the reconstructed model needs to be expanded in the axial direction, so that the estimated local SAR is closer to the real situation. At present, this aspect There is no corresponding research

Images