The turtle shell pattern matching seems like an interesting, but roughly solvable problem - don't even need anything as complex as a NNet or Deep learning, an S

The turtle shell pattern matching seems like an interesting, but roughly solvable problem - don't even need anything as complex as a NNet or Deep learning, an SVM, or heck even a well trained correlation filter (UOTSDF, MMCF, ...maybe even MOSSE) can be used after first registering the image (DNets maybe). It's nice to see a Detroit based tech activity that isn't recruiting (most end up in Ann Arbor). Too bad i'm pretty busy this weekend or i'd join in.