-
Notifications
You must be signed in to change notification settings - Fork 102
Description
- TIA Toolbox version: latest from the develop branch
- Python version: 3.8
- Operating System: Ubuntu 20.04
I have a batch of ome.tiff that are not compatible with slide_thumbnail(). Attempting to read at any level causes MemoryError. It looks like it's trying to read the base level array. The work around is to use read_bounds(), and specify the resolution=top level. For instance, a sample ome.tiff has the following dimensions:
[157521 161047]
[78760 80523]
[39380 40261]
[19690 20131]
[ 9845 10065]
This doesn't work:
thumbnail = wsi_reader.slide_thumbnail(resolution=4, units="level")
plt.imshow(thumbnail)
I get MemoryError: Unable to allocate 70.9 GiB for an array with shape (161051, 157525, 3) and data type uint8
This works:
target_level = wsi_reader_v1.info.level_count - 1
bounds = [0, 0, *wsi_reader_v1.info.level_dimensions[-1]]
thumbnail = wsi_reader.read_bounds(
bounds,
resolution=target_level,
units='level'
)
However, the resulting thumbnail has shape (629, 615, 3). Is that expected behavior?
Link to sample image: http://gofile.me/6VSoS/3cPUTcMiw
Update: Interestingly, if I specify coord_space = 'resolution' in read_bounds(), then I can reproduce the same memory error. Looks like that's where the bug is.