I have cleaned up the Houston Flooding segmenation project’s GitHub page a bit to make everything more presentable.

Also:

Here’s a link to a presentation that largely explains the process behind the model

(and there are a bunch of ‘appendix’ slides with extra statistics, pictures, ideas, etc.)

Please go take a look and/or contact if me if you are interested in learning more. I am always happy to discuss the project, it’s limitations, my burning ideas on how to overcome them, or anything else data science/imagery related.

And just for fun, here’s another ‘bonus’ segmenation image: