Recognizing temporal trajectories using the condensation algorithm

The recognition of human gestures in image sequences is an important and challenging problem that enables a host of human-computer interaction applications. This paper describes an incremental recognition strategy that is an extension of the "Condensation" algorithm proposed by Isard and Blake (1996). Gestures are modeled as temporal trajectories of some estimated parameter over time (in this case velocity). The condensation algorithm is used to incrementally match the gesture models to the input data. The method is demonstrated with an example of an augmented office white-board in which a user makes simple hand gestures to grab regions of the board, print them, save them, etc.

[1]  Richard Szeliski,et al.  Image mosaicing for tele-reality applications , 1994, Proceedings of 1994 IEEE Workshop on Applications of Computer Vision.

[2]  Quentin Stafford-Fraser,et al.  BrightBoard: a video-augmented environment , 1996, CHI '96.

[3]  Michael Isard,et al.  Contour Tracking by Stochastic Propagation of Conditional Density , 1996, ECCV.

[4]  James L. Crowley,et al.  Coordination of perceptual processes for computer mediated communication , 1996, Proceedings of the Second International Conference on Automatic Face and Gesture Recognition.

[5]  Hiroshi Ishii,et al.  Tangible bits: towards seamless interfaces between people, bits and atoms , 1997, CHI.

[6]  Michael Isard,et al.  A mixed-state condensation tracker with automatic model-switching , 1998, Sixth International Conference on Computer Vision (IEEE Cat. No.98CH36271).

[7]  Michael J. Black,et al.  Parameterized Modeling and Recognition of Activities , 1999, Comput. Vis. Image Underst..