From: Jerry Mouse on
Hello,

you have Job where you CAN work on watch recognizing ? REALLY ?

WOU WOU .... send me address of such company .... I WANT this JOB too .....

PLEASE ...PLEASE ...PLEASE ...PLEASE ...PLEASE ...PLEASE ...PLEASE ...PLEASE ...

JERRY

PS: Of course I have solution, original, working, easy.........

:-)
From: Jerry Mouse on
MY EMAIL.... jerry.mouse.email(a)seznam.cz
From: ImageAnalyst on
You didn't answer my question, but now I have another one. What kind
of company do you and Julia work for that needs to know what the time
on watches that are in advertisements is? Well I can tell you the
answer right now. It is ALWAYS some time between 10:08 and 10:10 -
never fail, ever. That should be accurate enough, because if it's not
- you're screwed. Those images are horrible. It's not impossible but
I'm not going to spend hours developing a solution for this. You
might try median filtering, morphological operations, houghlines,
thresholding and segmenting based on shape, etc. until you find
something that kind of works.
First  |  Prev  | 
Pages: 1 2 3
Prev: Video mosaicing
Next: break/return upon user input