Thank you for the question. The aim of this project is to return the post-deformation image to its original state in the pre-deformation image by coupled filtering method so as you can see in the results, the third image is the image after this process and the intensity of the tissue scatterers is closer to the pre-deform image, meaning that the process successfully compensate for some of the tissue movements. However, the correlation coefficient we got for the pre-deform and the output image is not that high so we still have room for improvement. And as a non-specialist, if you compare the original post-deform image with the outputted image, you can see which part of the tissue has undergone movements. I hope this answers your question.
To add on, one of our future developments is to highlight the differences between pre- and post- deformation images, and tells the user what movement (stretching, rotation, twisting) was estimated to be undergone within the tissue. In order to achieve this, however, we’ll need a much comprehensive dataset instead of 2D to visualize the tissue movement in a 3D coordinate plane.
As a non-specialist, what can we get from the output? Is the tissue in good shape?
Thank you for the question. The aim of this project is to return the post-deformation image to its original state in the pre-deformation image by coupled filtering method so as you can see in the results, the third image is the image after this process and the intensity of the tissue scatterers is closer to the pre-deform image, meaning that the process successfully compensate for some of the tissue movements. However, the correlation coefficient we got for the pre-deform and the output image is not that high so we still have room for improvement. And as a non-specialist, if you compare the original post-deform image with the outputted image, you can see which part of the tissue has undergone movements. I hope this answers your question.
To add on, one of our future developments is to highlight the differences between pre- and post- deformation images, and tells the user what movement (stretching, rotation, twisting) was estimated to be undergone within the tissue. In order to achieve this, however, we’ll need a much comprehensive dataset instead of 2D to visualize the tissue movement in a 3D coordinate plane.
We hope this answers your question!