1

I'm trying to create a network that does face-detection and also shows a persons age. displaying age is a regression problem.
Would adding a new parameter to the normal detection label suffice?
I mean currently this is what we do for a simple detection task : 

filename,width,height,class,xmin,ymin,xmax,ymax
resized_pic_0317.jpg,410,256,man,2,2,406,252

Would this be OK? :

filename,width,height,class,xmin,ymin,xmax,ymax,age
resized_pic_0317.jpg,410,256,man,2,2,406,252,58

if not, what would be the right sequence of actions to do?

4

0 回答 0