The code does not presently export in a human-friendly format where in the original video the ROIs correspond to. This should be done.
The current thinking is to take the first frame of the video and draw a rectangle denoting each RoI, and then draw the letter inside the rectangle. This way, a person can look at an image to check which bridge is which RoI.
JarredAllen
commented
5 years ago 
The code does not presently export in a human-friendly format where in the original video the ROIs correspond to. This should be done.
The current thinking is to take the first frame of the video and draw a rectangle denoting each RoI, and then draw the letter inside the rectangle. This way, a person can look at an image to check which bridge is which RoI.