How do you improve customisation when working with Computer Vision?

[alfie_36l4]

I have a question about Computer Vision that may be useful for other people working with it too.

I am particularly interested in customisation and testing. I would like to compare approaches that people have actually used rather than just theoretical suggestions.

Are there any settings, tools, or workflow changes that made a noticeable difference?