M. Turk. Visual interaction with lifelike characters. In Proc. Second IEEE Conference on Face and Gesture Recognition, Killington, October 1996.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....interaction, education and tutoring, smart rooms and cars, and surveillance systems. Many researchers are developing useful techniques for the rapidly growing area of multimodal and multimedia interaction. Turk s work at Microsoft with visual control of virtual actors is one example [71], as is Blake s use of sophisticated multi object trackers [72] in hand tracking [73] and new work by Frey and Jojic on learning exible sprites [4] Another well known centre for such work is the MIT Media lab led by Pentland. Some of this research has been mentioned in the reviews earlier, ....

M. Turk, \Visual interaction with lifelike characters," in IEEE International Conference on Automatic Face and Gesture Recognition, Killington, VE, 1996.

....model matching. Ongoing research at the MIT Media Lab has shown progress in the modelling and interpretation of human body activity [22, 28, 29] Computationally simple view based approaches to action recognition have also been proposed [4] and similar attempts have been made at Microsoft Research [26, 6]. However, these systems do not attempt intentional tracking and modelling to control active cameras for Visually Mediated Interaction (VMI) Previous work on vision based camera control has been based on off line execution of pre written scripts of a set of defined camera actions [23] This ....

M. Turk. Visual interaction with lifelike characters. In Proceedings of International Conference on Automatic Face & Gesture Recognition, pages 368--373, Killington, VT, 1996. IEEE Computer Society Press.

....recalling some of the past interactions. It would certainly be useful to determine who is talking to whom in case of multiple participants. Vision can also provide speech recognition with a more accurate input in a noisy environment by focusing the attention to the spatial location of the user [80]. This is achieved either by a post filtering step when using a phased array of microphones or, more actively, by directing a parabolic microphone to the intended source. Finally, vision can also prove helpful for phoneme disambiguation, i.e. lip reading. An important application area in the user ....

....prove helpful for phoneme disambiguation, i.e. lip reading. An important application area in the user interface domain involves social interfaces. Social interfaces deal with computer generated characters, with human like behaviors, who attempt to interact with users in a more personable way [80]. Alternative application areas in the user interface domain are sign language translation, gesture driven control of graphical objects or appliances, and signaling in high noise environments such as factories or airports. In the motion analysis domain, a possible application is content based ....

[Article contains additional citation context not shown here]

M. Turk. Visual interaction with lifelike characters. In Proc. of IEEE International Conference on Automatic Face and Gesture Recognition, pages 368--373, Killington, 1996.

....recalling some of the past interactions. It would certainly be useful to determine who is talking to whom in case of multiple participants. Vision can also provide speech recognition with a more accurate input in a noisy environment by focusing the attention to the spatial location of the user [80]. This is achieved either by a postfiltering step when using a phased array of microphones or, more actively, by directing a parabolic microphone to the intended source. Finally, vision can also prove helpful for phoneme disambiguation, i.e. lip reading. An important application area in the user ....

....also prove helpful for phoneme disambiguation, i.e. lip reading. An important application area in the user interface domain involves social interfaces. Social interfaces deal with computergenerated characters, with human like behaviors, who attempt to interact with users in a more personable way [80]. Alternative application areas in the user interface domain are sign language translation, gesture driven control of graphical objects or applicances, and signaling in high noise environments such as factories or airports. In the motion analysis domain, a possible application is content based ....

[Article contains additional citation context not shown here]

M. Turk, Visual interaction with lifelike characters, in Proc. of IEEE International Conference on Automatic Face and Gesture Recognition, Killington, 1996, pp. 368--373.

....model matching. Ongoing research at the MIT Media Lab has shown progress in the modelling and interpretation of human body activity [24, 30, 31] Computationally simple view based approaches to action recognition have also been proposed [4] and similar attempts have been made at Microsoft Research [28, 6]. However, these systems do not attempt intentional tracking and modelling to control active cameras for Visually Mediated Interaction (VMI) Previous work on vision based camera control has been based on off line execution of pre written scripts of a set of defined camera actions [25] This ....

M. Turk. Visual interaction with lifelike characters. In Proceedings of International Conference on Automatic Face & Gesture Recognition, pages 368--373, Killington, VT, 1996. IEEE Computer Society Press.

....model matching. Ongoing research at the MIT Media Lab has shown progress in the modelling and interpretation of human body activity [25, 31, 32] Computationally simple view based approaches to action recognition have also been proposed [4] and similar attempts have been made at Microsoft Research [29, 6]. However, these systems do not attempt intentional tracking and modelling to control active cameras for Visually Mediated Interaction (VMI) Previous work on vision based camera control has been based on off line execution of pre written scripts of a set of defined camera actions [26] This ....

M. Turk. Visual interaction with lifelike characters. In Proceedings of International Conference on Automatic Face & Gesture Recognition, pages 368--373, Killington, VT, 1996. IEEE Computer Society Press.

....s#i# # # v i j2x i , x p#i# , x s#i# j : Here, i indexes a predetermined number of probes (16 are used for face tracking) x i is Probe i s pixel location, v i is Probe i s expansion direction, and p and s return the predecessor and successor indices of i. The algorithm is similar to draping [22] and to balloons [4] but it is more efficient, and it avoids spurious edges when tracking faces by constraining shape to star shaped polygons. Approximate position in the plane parallel to the camera image is tracked. Layer 4 is the same as Layer 3, with an additional computation of the principal ....

M. Turk. Visual interaction with lifelike characters. In Proc. Automatic Face and Gesture Recognition, 1996.

No context found.

M. Turk, "Visual interaction with lifelike characters," Proc. Second IEEE Conference on Face and Gesture Recognition, Killington, VT, October 1996.

....and track the subject s body using an articulated kinematic model (Section 6) 4. Fast, Simple Head Tracking In this section we present a simple but fast technique to track a user sitting at a workstation, locate his head, and use this information for subsequent gesture and pose analysis (see [14] for more details) The technique is appropriate when there is a static background and a single user a common scenario. First a representation of the background is acquired, by capturing several frames and calculating the color mean and covariance matrix at every pixel. Then, as live video ....

M. Turk, "Visual interaction with lifelike characters," Proc. Second IEEE Conference on Face and Gesture Recognition, Killington, VT, October 1996.

No context found.

M. Turk. Visual interaction with lifelike characters. In Proc. Second IEEE Conference on Face and Gesture Recognition, Killington, October 1996.

No context found.

M. Turk. Visual interaction with lifelike characters. In Proc. Automatic Face and Gesture Recognition, 1996.

No context found.

M. Turk. Visual interaction with lifelike characters. In Proc. Second IEEE Conference on Face and Gesture Recognition, Killington, October 1996.

No context found.

M. Turk: "Visual interaction with lifelike characters", IEEE computer society press, October 1996.

No context found.

M. Turk. Visual Interaction with Lifelike Characters. In International Conference on Automatic Face and Gesture Recognition, Killington, VT, USA, 1996.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC