Memory efficient
- max scale clipping to 10 is omitted, compute scale for rish only
- separate ncpu parameter introduced for ants
- statistics are written to a text file for future use
- print a msg when --travelHeads is triggered
- decouple shm_coeff from spm_bspline computation to avert MemoryError
- decouple dti and rish computation for increasing efficiency
- Use Popen() and wait() method to instead of check_call to avert MemoryError
- remove nan in GFA image