Toward extraction with deep nets
This whole writeup is based on experiences with what you might call “old school” image processing, in which the human comes up with an entirely heuristic model to achieve their objective.
I’m curious to find out how it holds up as a way to approach image extraction problems using deep learning, which tends to favor a more data-heavy / heuristics-light approach.
If you have opinions.. let me know!
Note: we’re hiring!
If you like solving puzzles like this and want to make a big impact in industry, reach out and say hello. We’re building an amazing team at Instabase and looking to grow.