An Image Registration Method Based on Improved Optical Flow Model
An image registration and optical flow field technology, applied in the field of computer vision, can solve the problems of loss of details and insufficient precision, and achieve the effects of protecting edge features, strong robustness, and avoiding over-smoothing
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0035] An image registration method based on an improved optical flow field model, see figure 1 , the method includes the following steps:
[0036] 101: Construct the energy function of the optical flow field model composed of data items, anisotropic regular items, and non-local smooth items;
[0037] 102: Solve the displacement field by minimizing the energy function, and calculate the displacement field for each layer image of the pyramid, and use the displacement field of each layer as the initial displacement field of the next layer until the final displacement field is obtained;
[0038] 103: Perform change interpolation on the image to be registered according to the obtained final displacement field, to obtain a final registration image, and realize registration.
[0039] To sum up, through the above steps 101-103, the embodiment of the present invention improves the optical flow model for the problems caused by the traditional optical flow model, such as over-smoothing...
Embodiment 2
[0041] The scheme in embodiment 1 is further introduced below in conjunction with specific calculation formulas and examples, see the following description for details:
[0042] 201: Construct the energy function of the optical flow field model;
[0043] The data items in the traditional optical flow model are in the square form, which will amplify the displacement estimation difference of the overflow point. In order to increase the penalty for the overflow point, a non-square penalty function is used. The data item is defined as follows:
[0044] E. D =∫ Ω ψ(|I 2 (X+W)-I 1 (X)| 2 )dX (1)
[0045] In the formula, I 1 and I 2 are the 2 images to be registered; X=(x,y) T Represents a point in the image space domain Ω; W=(u,v) T (u and v are the optical flow horizontal displacement and vertical displacement respectively) represent the image I 1 and I 2 The motion displacement field between.
[0046] The traditional optical flow field model uses an isotropic regular...
Embodiment 3
[0074] The technical solutions of the present invention will be further described in detail below in conjunction with specific examples.
[0075] Figure 2-Figure 4 It is a schematic diagram of the comparison between the registration results of this method and the traditional optical flow field model algorithm. figure 2 is the medical MRI (magnetic resonance imaging) image registration result map, from figure 2 It can be seen that the traditional H-S algorithm has a certain correction effect on the image, but the overall boundary of the image is blurred, there are many noise points, and the registration effect is not ideal; compared with it, the registration effect of the Brox algorithm is greatly improved, and the structure remains relatively Complete, but the correction effect of the details in some small displacement areas is not ideal; there are obvious breakpoints in the registration results of SIFTFlow, and there are obvious block effects because the registration accu...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More