Hello TDers, I’d like to start to let my computer recognise objects, eventually predict their positions and finally using the positions to create events.
Would this be possible with TouchDesigner?
Some friends adviced me to use neural networks and especially tensorflow. Can anyone point me how to start with it (from the installation to the first steps)?
Thank you!
You can definitely do that in TD, but I’d only use it if you want to be doing things in TD anyway. You’ll have to write a fair amount of Python glue code to use neural networks in TD. I’ve been doing similar with a lot of facial recognition in the last few months, but there’s more than a little setup involved to make sure you’ve installed all the prerequisites and set TD’s environment up to load them all. If you’re experienced with Python, it’s doable. If Python is new to you as well… it could be kind of a long road.
I intend to open source some of the setup work I’ve done, but I’ve got to clear that with my employer first, so that’s a little while off.
thank you @Tyrell @evan_jones !
sounds very complicated at the moment for my knowledge considering that I have just started using easy scripts with python since a couple of months.
However I will have a look to some general tutorials on how tensorflow works.