An Evaluation of Tracking Methods for Human-Computer Interaction

Tracking methods are evaluated in a real-time feature tracking system used for humancomputer interaction (HCI). The Camera Mouse, a HCI system that uses video input to manipulate the mouse cursor, was used as the test platform for this study. The Camera Mouse was developed to assist individuals with severe disabilities in using computers, but the technology may be used as a method for HCI for those without disabilities as well. The initial location of the feature to be tracked is set manually by the user by clicking on a portion of the image window and the system tracks its location in subsequent frames. Tracking methods were evaluated in terms of the best combination of computational efficiency and accuracy. Trackers evaluated include feature matching using normalized correlation coefficients and optical flow, in the form of a Lucas-Kanade tracker. Each tracking algorithm is evaluated both with and without the addition of Kalman Filters. The effects of 2-D (x and y location) 4-D(location and velocity in the x and y directions), and 6-D (location, velocity, and acceleration in the x and y directions) are examined and analyzed. The normalized correlation coefficient tracker, without Kalman Filtering, was found to be the best tracker in terms of suitability to the human-computer interaction system implemented.

[1]  James M. Rehg,et al.  A multiple hypothesis approach to figure tracking , 1999, Proceedings. 1999 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No PR00149).

[2]  Markus Kohler,et al.  Using the Kalman Filter to track Human Interactive Motion - Modelling and Initialization of the Kalm , 1997 .

[3]  Margrit Betke,et al.  THE CAMERA MOUSE: PRELIMINARY INVESTIGATION OF AUTOMATED VISUAL TRACKING FOR COMPUTER ACCESS , 2000 .

[4]  E. Brookner.  Tracking and Kalman Filtering Made Easy , 1998 .

[5]  Takeo Kanade,et al.  An Iterative Image Registration Technique with an Application to Stereo Vision , 1981, IJCAI.

[6]  Massimo Tistarelli,et al.  Multiple Constraints for Optical Flow , 1994, ECCV.

[7]  Hans-Hellmut Nagel,et al.  Optical Flow Estimation: Advances and Comparisons , 1994, ECCV.

[8]  M. Betke,et al.  The Camera Mouse: visual tracking of body features to provide computer access for people with severe disabilities , 2002, IEEE Transactions on Neural Systems and Rehabilitation Engineering.

[9]  Thierry Pun,et al.  Feature selection for object tracking in traffic scenes , 1995, Other Conferences.

[10]  T. Başar,et al.  A New Approach to Linear Filtering and Prediction Problems , 2001 .

[11]  Berthold K. P. Horn,et al.  Determining Optical Flow , 1981, Other Conferences.

[12]  Takeo Kanade,et al.  Shape and motion from image streams under orthography: a factorization method , 1992, International Journal of Computer Vision.

[13]  G.E. Moore,et al.  Cramming More Components Onto Integrated Circuits , 1998, Proceedings of the IEEE.

[14]  David A. Landgrebe,et al.  Covariance Matrix Estimation and Classification With Limited Training Data , 1996, IEEE Trans. Pattern Anal. Mach. Intell..

[15]  Trevor Darrell,et al.  Reducing drift in parametric motion tracking , 2001, Proceedings Eighth IEEE International Conference on Computer Vision. ICCV 2001.

[16]  Wouter Pasman,et al.  Alternatives for optical tracking , 2001 .