Browse Source

Fix dtype at np.float64

master
Ruben Verweij 8 years ago
parent
commit
a56a7cc665
1 changed files with 5 additions and 9 deletions
  1. +5
    -9
      nd2reader/parser.py

+ 5
- 9
nd2reader/parser.py View File

@ -94,19 +94,15 @@ class Parser(object):
else: else:
return raw_image_data return raw_image_data
def get_dtype_from_metadata(self):
@staticmethod
def get_dtype_from_metadata():
""" """
Determine the data type from the metadata. Determine the data type from the metadata.
For now, always use float64 to prevent unexpected overflow errors when manipulating the data (calculating sums/
means/etc.)
:return: :return:
""" """
bit_depth = self._raw_metadata.image_attributes[six.b('SLxImageAttributes')][six.b('uiBpcInMemory')]
dtype = np.float64
if bit_depth <= 16:
dtype = np.float16
elif bit_depth <= 32:
dtype = np.float32
return dtype
return np.float64
def _check_version_supported(self): def _check_version_supported(self):
""" """


Loading…
Cancel
Save