Novel non-rigid image registration method
An image registration, non-rigid technology, applied in the field of image processing, can solve unreasonable problems and achieve effective registration effect
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
example 1
[0104] This example is to register a pair of artificial images, such as figure 2 As shown, (a) is a fixed image, (b) is a floating image, (c) is the DiffeoDemon model registration result, (d) is the VTV model registration result, (e) is the BD registration result, (f) is Registration results of the BGD model of the present invention.
[0105] From figure 1 As can be seen in the figure, the registration result obtained by the DiffeoDemons model is obviously too smooth, and the smear effect is relatively heavy. The registration results of the VTV model have obvious jagged edges. This example shows that the BGD model can register images with sharp edges and large grayscale variations.
[0106] Furthermore, in order to compare different non-rigid registration models, we use some well-recognized and effective evaluation indicators to evaluate the registration results of each model. Here we use the classic average structural similarity mSSIM, normalized mutual information (NMI)...
example 2
[0111] In this example, the non-rigid medical image registration BGDSSD model based on the bounded generalized deformation function is applied to the registration of liver CT images, such as image 3 As shown, (a) is a fixed image, (b) is a floating image, (c) is the DiffeoDemon model registration result, (d) is the VTV model registration result, (e) is the BD registration result, (f) is Registration results of the BGD model of the present invention.
[0112] Table 2 is from the average structure similarity mSSIM, normalized mutual information (NMI) and normalized cross-correlation coefficient (NCC) three quantitative indexes for embodiment 1 respectively in DiffeoDemons model, VTV model, BD model and the present invention The registration results under the BGD model are compared, and the optimal value is bolded. Table 2 illustrates the effectiveness of the BGD model of the present invention.
[0113] Table 2 Quantitative indicators of liver CT image registration results
...
example 3
[0116] This example tests the BGD registration model of the present invention on two 3-dimensional public datasets. These two datasets are 3D lung CT image sequences, called 4D-CT and COPDgene datasets, respectively.
[0117] The 4D-CT dataset contains lung CT images of 10 different test subjects, and each test subject has 3D volume data at 10 different moments. Along the x-y-z direction, the average voxel resolution of the volume data of the first 5 testers is 1.1×1.1×2.5mm 3 , the average number of voxels is 256×256×103. Along the x-y-z direction, the average voxel resolution of the 3D volume data of the last 5 testers is 0.97×0.97×2.5mm 3 , the average number of voxels is 512×512×128 (see https: / / www.dir-lab.com / ReferenceData.html for detailed introduction). The 3D image in the exhalation phase and the 3D image in the inhalation phase are respectively used as a fixed image and a floating image. The image grayscale of this dataset is roughly in the range of [0,4000]. Th...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More