Dr image distortion correction

Dr images are prone to S-shape, pillow, and local distortion, which interferes with the doctor's image judgment. In addition, in the three-dimensional reconstruction of medical images, uncorrected images can be reconstructed, which may cause interference such as ghosting. Therefore, distortion correction is a problem that must be addressed after Dr images are processed.

Figure 1 original target image

From figure 1, we can see that there is obvious pillow distortion near the image edge, and there is a certain degree of s distortion in the horizontal direction. The first step of correction is to mark the center of all circles in the image and create a standard grid distribution, as shown in effect 2. It should be noted that the accuracy of the center positioning of the circular target will greatly affect the accuracy of subsequent calibration. Here, the author also adds angle judgment to enable Adaptive Calibration and correction.

Figure 2 Circle Target Positioning Result and standard grid matching

CreateCubeSolve the calibration parameters. Finally, the bilinear difference value is used to obtain the corrected effect, as shown in figure 3.

Figure 3 corrected Image

We drew standard gridlines on the corrected image and found that the correction was perfect. After enlarging the image, you will find that the distorted circles (after distortion, they are elliptical) on the edge are also corrected as circles. This parameter can be used to correct other images. Figure 4 is the source image, with obvious distortion at the edge. Figure 5 after correction disappears.

Figure 4 Original to-be-corrected figure 5 distortion correction effect