a Readvtk matlab code was provided to read text files with specific format. The data is stored in array (3, NumberOfPoint) which is 3-D euclidean numeric data. Readvtk is a very simple Matlab code with less than 30 lines, actually around 10 lines are functioning.
b Writevtk matlab code was provided to write back the processed data to the specific format. So the processed data must be in the format that can be ran by Writevtk. Writevtk is also a very simple Matlab code with around 40 lines of code.
c 500 files contain a few hundreds points (3-D x,y,z axis) to a few hundreds thousand points for TESTING purpose. These files are already clustered.
a In order to test the program, a merge program needed to be written to merge the above 200 files (that mixed the data), the original format need to be keep so that it can be view by Paraview or Slicer. (I don't expect more that 30-40 command lines. You can make use of Readvtk to simply the code)
b Write a file to cluster (Hierachical - simple linkage clusting) these merged data into fiber (instead of sphere usually applied academically). Preferably you should modulate the code so that it can modify to K-mean clustering and other clustering.
As the merged data would contain more that 1M points of data, I cannot accept algorithms with running speed proportion to over to the power (1.5) of data size. Working principle like quicksort of mergesort would be OK. (I expected less than 60 command lines by applying recuring looping function).
If you are interested and need more information, please reply before today noon by email. Please give me a quote.( approximate price and working hours)