Igneous, our Python software for managing Neuroglancer 3D images, meshes, and skeletons, is now published open access in Frontiers in Neural Circuits!
Aleksandar Zlateski @jalex_bae@tartavull@nkemnitz1@jingpeng_wu@SebastianSeung
https://t.co/crntvRZqVW
cloud-files 6.6.0
- General transfers (e.g. gcs to s3) are twice as fast (switched from download/upload cycles to producer/consumer threads)
https://t.co/q0qud4VPdH
Popular political platform: humans should not cede even the tiniest fraction of control to machines, no matter how "intelligent." And those who are working to reduce human control to be stopped at all costs.
crackle-codec 0.53.0
- Query if a set of points are inside or outside a mask
- Compute the distance transform of sparse points for arbitrarily large shapes
- Fix for point clouds
- Parallelized contacts (accelerates hold filling in fastmorph)
https://t.co/sPEvE9mmaN
gaara 0.13.0
- Adds support for CLI
- Adds support for a voxel connectivity graph (binary images only)
- Fixes some (not all) bugs with chunked processing
https://t.co/lshKYEsXAy
@WillCJohns@HoffProf This would not work as there are "humanizers" that can be employed to hack the pangram score. I don't know how their results read to a human though.
Cool new feature in crackle for handling larger-than-RAM segmentations.
You can now apply infix operators like a+b, a==b, a!=b to process larger than RAM volumes. Supported operators: + - * // % ^ & << >> | == !=
Also:
result = arr.apply_binary_mask(mask)
Let me give a small example. Say you have two non-overlapping labels (maybe from different fluorophores) that are 50 GVx each and want to combine them into a single image and then apply an ROI mask.
combined = img1 + img2
result = combined.apply_binary_mask(mask)
I also added support to fastmorph for dilation and erosion on larger-than-RAM images.
It's true that this approach doesn't extend infinitely, but given crackle compression ratios between 6 (horrible L2 images) to 100,000 (a binary mask), you can get away with quite a lot.