I had no issues with Posenet, very well documented and easy to use, it worked very well on security camera images that had a "flat" enough viewing angle, a sample video I made:
I passed my modified Posenet sample code on to some of my pre-retirement colleagues who study "gait", their holy grail is "marker-less" body segment trajectories.
Obviously hollywood might use this technology someday for their green screen character animations if they aren't already.
We do get off topic here some times, but node-red brings together such a diverse and competent community that I always seem to learn something useful every time!
"We do get off topic here some times, but node-red brings together such a diverse and competent community that I always seem to learn something useful every time!"
do you have your code for posenet experiments sowhere? github?
i experimented with your ai-detector code for my nodered, was great!
may i try something with posenet from your code...? or do i have to retire beforhand?