We have reviewed others’ codes in our project in order to work with them, and figure out the basic morphing method used in this project.
Actually, in the previous blog, I have learned the Mesh Warping. But to apply this method into infrared image processing for medical using, there is still a long way to go.
We use OpenCV (C++ interface) to put up a framework. Other students have already finished the work about stiching, binarization, and dividing the image into different segments based on medical reserach.
And our work is to morph one segment to another standard segment we’ve got in previous work.
I’m still new to C++, so I wrote such a class to make the further work easier and clearer.
Mark and Get the Mesh
When we have 2 images to morph, we need to get the control points’ coordinates(x, y). And in an easiest way, we just sweep the image linearly to get the mesh.
Get the binary image of this segment.
Sweep every row, find the start point and the end point.
Put control points uniformly in the row.
The Control Points
This image shows a binary image for the standard leg.
Here is the morphing result. There are still some edges and corners morphed not so good.