November 19, 2006

Up todate code that works pretty well

The code: Download

So far the code displays pictures that have been filtered by the blob and then the contourProcessor applied. WIth the model overlayed in red, which is what the model percieves to be the person.(Hoi correct if im wrong)

currently all it does it display this and does not use it in a useful way, some means of calculating what direction the person is heading in needs to be created. To give the robot as well as means of grabbing the images into CImg format.

but other than that we are getting there.

Also talking to Matt in CS, said that the lists in cpp are very inefficient and it might be worth converting them back to vectors if performance is too slow.


- One comment Not publicly viewable

  1. Found a bug in contour processor – the shiftmodel method doesn’t work because you’re copying the pixel object and then changing it. Fixed it, I’ll post it soon.

    20 Nov 2006, 02:24

Add a comment

You are not allowed to comment on this entry as it has restricted commenting permissions.

November 2006

Mo Tu We Th Fr Sa Su
Oct |  Today  | Dec
      1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18 19
20 21 22 23 24 25 26
27 28 29 30         

Search this blog



Most recent comments

  • The reflective writing is submitted individually so make sure you submit it yourself tomorrow mornin… by on this entry
  • Yeh Ant's charts look great. Not sure – Ant reckons 1650 max (10% over), so if anyone has any sugges… by on this entry
  • Wow, the report looks really good :) Do you think it will matter too much if it is a couple of hundr… by on this entry
  • nice, overs better than under ;) by on this entry
  • I'll do the other bits then: introduction/overview of person tracking, testing/problems (if Ant+Hoi … by Matt on this entry

Blog archive

Not signed in
Sign in

Powered by BlogBuilder