source: level2/competencies/README.txt @ 157

Revision 157, 1.8 KB checked in by ginevra, 10 years ago (diff)
Line 
1{\rtf1\ansi\ansicpg1252\deff0\deflang2057{\fonttbl{\f0\fnil\fcharset0 Courier New;}{\f1\fnil\fprq1\fcharset0 Courier New;}{\f2\fswiss\fcharset0 Arial;}}
2{\*\generator Msftedit 5.41.21.2508;}\viewkind4\uc1\pard\f0\fs20 ************************************\par
3FaceTracking\par
4************************************\par
5\par
6This code allows for the detection and tracking of a face in the scene. Face detection is performed using Haar classifiers and is based on the OpenCV face detection code.\par
7Tracking is performed using a Camshift wrapper (see License.txt, in the Camshift wrapper folder, for more details) based on the Camshift algorithm provided by OpenCV, which tracks a combination of colours.\par
8\par
9The main program waits until a face is detected in the scene (using the function \f1 waitForFaceDetect(\f0 )); when a face is detected, the tracking is automatically initialised using the face bounding box returned by the Haar classifier.\par
10\par
11You may need to modify the paths for the Haar classifiers (in the function \f1 InitFaceDetection() of the class FaceDetection\f0 ) to reflect your directory structure.\par
12\par
13The functions setVmin() and setSmin() allow for the setting of the Camshift parameters. The suitable values to use may change according to your application and setup. \par
14\par
15The program allows for a re-initialisation of the tracking.\par
16The user can re-initialise the tracking process by pressing the 'r' key. When the 'r' key is pressed, the program waits until a face is detected and re-initialise the tracker using the new face bounding box.\par
17\par
18The code has been written and tested in Windows, but it should work under Linux as well.\par
19\par
20You will need to install OpenCV on your machine.\par
21\par
22Questions to: ginevra@dcs.qmul.ac.uk\f2\par
23}
24
Note: See TracBrowser for help on using the repository browser.