production code for wasp sample 3 47_seg_v5

• rewrite the disjoint set from scratch
• disjoint set serialization works but there is an unicode issue to load in python

production code for wasp sample 3 47_seg_v2_splitted

v0.0.9

comment the print

seeded watershed

• seeded watershed to dilate objects in segmentation with the guidance of affinity map.

detect T-bars; improve preprocessing.

v0.0.7

detect tbar points

used in production to segment wasp sample 1 (53_seg)

• split big mergers by deleting the corresponding edges in dendrogram.

used in production to segment wasp sample 1 (52_seg)

v0.0.5

const reference to make it a little bit faster

global segmentation ID remapping

• serialization of dendrogram, make it pickable

watershed and mean affinity agglomeration.

both work great!

separate modules using cmake to compile

• compile the modules separately using cmake to speed up development.
• currently, we have two modules: skeleton, segmentation

skeleton and nblast

• downsample
• nblast