Join devRant
Do all the things like
++ or -- rants, post your own rants, comment on others' rants and build your customized dev avatar
Sign Up
Pipeless API
From the creators of devRant, Pipeless lets you power real-time personalized recommendations and activity feeds using a simple API
Learn More
-
You could script that in blender providing you can make/rig the model (or have one already).
Actually theres plenty of footage on YouTube. One good idea would be to compile snippets of people from different angles and different activities as a sort of training database
The angles data labels are the most important though. -
@Wisecrack well it would be more about finding people than labeling angles although come to think of it that is a pretty good idea.
a person of this girth from this euclidean angle and such is probably gonna look pretty standard when wearing these clothes or in these conditions.
you could also simulate airflow.
that would be 360^3 output neurons+ however.
also at least one to say 'THIS IS PERSON !'
or is it ?
don't know much about nn... can you output real number values that have meaning ?
People are like a stable set of database artifacts blah blah
Repeat this repost that
Did anyone ever figure out why it is that trained human detection models don’t work better ?
All you’d think they’d have to do is run an animation renderer to create data representing a human figure in every conceivable position imaginable with limbs moving into positions that were within certain tolerances in different positions with different textures
I don’t personally think the full capability is being represented
random