Open praveenivp opened 2 years ago
When splitting the data in cluster. the MFI weights calculation suffers from numerical errors due to poor scaling. The volumes reconstructed from different node has different intensity distribution. Splitting the matrix is probably not a good idea.
When splitting the data in cluster. the MFI weights calculation suffers from numerical errors due to poor scaling. The volumes reconstructed from different node has different intensity distribution. Splitting the matrix is probably not a good idea.