Difference between revisions of "Face detection"

From AIRWiki
Jump to: navigation, search
m
 
(5 intermediate revisions by 2 users not shown)
Line 5: Line 5:
 
|coordinator=AndreaBonarini
 
|coordinator=AndreaBonarini
 
|tutor=AndreaBonarini;DavideRizzi
 
|tutor=AndreaBonarini;DavideRizzi
|students=GiulioFiscella; FedericoSem;  
+
|students=GiulioFiscella; FedericoSem;
 
|resarea=Robotics
 
|resarea=Robotics
 
|restopic=Robot development
 
|restopic=Robot development
 
|start=2010/04/15
 
|start=2010/04/15
|end=2010/08/31
+
|end=2010/08/30
|status=Active
+
|status=Closed
 
|level=Bs
 
|level=Bs
 
|type=Thesis
 
|type=Thesis
 
}}
 
}}
$wgUseImageMagick = true;
+
Aim of this project was the development of a vision tracking system to be implemented for [[E-2? - A robot for exhibitions | E-2?]] in order to detect and follow, in a robust way, faces in uncostrained enviroment (i.e. an exhibition or mall environment).
Aim of this project is the development of a system tobe used with [[E-2? - A robot for exhibitions | E-2?]] to detect in a robust way faces in an exhibition or mall environment.
+
  
This is a brief diagram which shows how our candidate algorithm work:
+
Our efforts have lead to an integration of different well known vision algorithms, implemented in C++ with the help of the opensource OpenCV libraries, which has proved to be much more reliable in the tracking tasks than the previous vision system, even if it is still affected by some detecting issues.
[[Image:Diagram.jpg|400px|thumb|left]]
+
 
 +
Here you can see a video showing the algorithm performance while working on a simple notebook webcam:
 +
 
 +
 
 +
 
 +
''video to be uploaded...''
 +
 
 +
 
 +
 
 +
Finally, this is a short video about a real time test on the robot.
 +
 
 +
 
 +
 
 +
''video to be uploaded...''
 +
 
 +
 
 +
 
 +
More details can be found in the discussion page.

Latest revision as of 14:20, 3 October 2011

Face detection
Error creating thumbnail: Unable to save thumbnail to destination
Short Description: Detect faces in a crowded environment
Coordinator: AndreaBonarini (andrea.bonarini@polimi.it)
Tutor: AndreaBonarini (andrea.bonarini@polimi.it), DavideRizzi ()
Collaborator:
Students: GiulioFiscella (giulio.fiscella@torrescalla.it), FedericoSem (fede_sem@tele2.it)
Research Area: Robotics
Research Topic: Robot development
Start: 2010/04/15
End: 2010/08/30
Status: Closed
Level: Bs
Type: Thesis

Aim of this project was the development of a vision tracking system to be implemented for E-2? in order to detect and follow, in a robust way, faces in uncostrained enviroment (i.e. an exhibition or mall environment).

Our efforts have lead to an integration of different well known vision algorithms, implemented in C++ with the help of the opensource OpenCV libraries, which has proved to be much more reliable in the tracking tasks than the previous vision system, even if it is still affected by some detecting issues.

Here you can see a video showing the algorithm performance while working on a simple notebook webcam:


video to be uploaded...


Finally, this is a short video about a real time test on the robot.


video to be uploaded...


More details can be found in the discussion page.