Fig 1.
Technical Roadmap.
Fig 2.
ASPP Basic Structure.
Fig 3.
Basic structure of MFAB module.
Fig 4.
Improve the basic structure of 3D U-Net network.
Table 1.
Change analysis of loss function before and after improvement of glioma segmentation model.
Table 2.
Analysis of Changes in Loss Function under Different Levels of Interference before and after Improvement.
Table 3.
Results of Jaccard Index Index under Different DataSets.
Table 4.
Performance Comparison and Analysis of Various Network Models.
Table 5.
Analysis of the influence of the selection of loss function on the segmentation accuracy of glioma.
Fig 5.
Original 3D-MRI brain glioma case sample data (a) The 1st 3D-MRI glioma sample (b) The 2nd case of 3D-MRI glioma (c) The 3rd case of 3D-MRI glioma (d) The 4th case of 3D-MRI glioma (e) The 5th case of 3D-MRI glioma (f) The 6th case of 3D-MRI glioma.
Fig 6.
Sample labeling results of 3D-MRI glioma cases (a) The 1st 3D-MRI glioma sample (b) The 2nd case of 3D-MRI glioma (c) The 3rd case of 3D-MRI glioma (d) The 4th case of 3D-MRI glioma (e) The 5th case of 3D-MRI glioma (f) The 6th case of 3D-MRI glioma.
Fig 7.
The segmentation results of 3D-MRI glioma case samples using the method described in this article (a) The 1st 3D-MRI glioma sample (b) The 2nd case of 3D-MRI glioma (c) The 3rd case of 3D-MRI glioma (d) The 4th case of 3D-MRI glioma (e) The 5th case of 3D-MRI glioma (f) The 6th case of 3D-MRI glioma.
Fig 8.
Sample segmentation results of 3D-MRI glioma cases using comparative methods (a) The 1st 3D-MRI glioma sample (b) The 2nd case of 3D-MRI glioma (c) The 3rd case of 3D-MRI glioma (d) The 4th case of 3D-MRI glioma (e) The 5th case of 3D-MRI glioma (f) The 6th case of 3D-MRI glioma.
Fig 9.
HD95 distance results.
Fig 10.
Segmentation Error Results.
Fig 11.
Linear time complexity results.