VS siCep55     (20231019-t1_siCep55-50-4_mitosis_21_25_to_117.tiff)
92 frames

pool: 1.736s
thread: 2.309s
np_thread: 1.706s

They all spawn 25 threads

thread >> np_thread; because custom splitting algorithm is worse than numpy's:
basically with custom algo they all get 3 frames except last thread that gets 17.
(longest processing time should be around 17 frames)
While numpy's algorithm gives 3/4th 4 frames and 3 frames to the last 1/4th;
(longest processing time should be around 4 frames).




VS example

np_thread: 
(49 frames)

pool: 1.758s
thread: 3.607s
np_thread: 1.781s


