This is a part of my Capstone Research. 

After the determination of feature points (turning points and radius sudden changing points) in a branch, the following method is used to match other common nodes in a branch.

First, get feature point indexes of two branch. For example, t1 branch is defined as base branch, t2 branch as matching branch. In a limited area, pair these feature nodes so that the branch is separated to a few sub-branches. If not paired successfully, it does not matter. The pairing process is used to enhance the accuracy but not an essential step.

In each sub-branch, I assume the voltage between two adjacent nodes is linear, so that proportion of length away from a node can be converted to a proportioned difference of voltage. Get the voltage of sub-branch t1 and t2. Resize and unify t2 according to t1, so that first node of t1 and t2 is the same, and last node of t1 and t2 is 0. 

Starting from second node to the second last node in t2, find two adjacent nodes in t1, meeting the requirement that one of the voltage is higher and the other is lower than the nodes in t2. Calculate the percentage of deviation to these two nodes, find a smaller one. Record the pair of node ids in sub-branches, percentage of deviation and its basing voltage(deviation of two adjacent nodes in t1)

Combine each pair of sub-branch to a whole pair of branch.

In painting the process of pairing t1 and t2, percentage of voltage deviation and its basing voltage are converted to location deviation to basing nodes.

In this graph, blue nodes are from t1 branch, red nodes are from t2 branch.

Detailed code: 

sub-branch pairing: findbranchipair.m

paint paired t1&t2 branch: paintcompbranch.m