Dark 2D Classes after Min over Scale

I am facing a similar issue of light vs dark 2D classes upon merging particles from two datasets, have just posted herein seeking for advice please:

hoping to understand what I might doing wrong or how to fix