| Sun, X., Foote, J., Kimber, D., and Manjunath, "Panoramic video capturing and compressed domain virtual camera control", ACM Multimedia, pp. 229-238, 2001. |
....The user will be able to navigate using VRML and access spatial locations. There have been intelligent cameras developed to follow the predefined objects. An intelligent camera system is developed to track lecturer and audience in [79] Panoramic cameras are used to capture the environment in [96, 53]. The region of interests are detected from the panoramic scene and these parts are presented [96] This approach is not applicable to previously developed videos. A spatially indexed camera is explained for navigation in [53] Our goal is to enable spatial browsing of recorded video. We do not ....
....cameras developed to follow the predefined objects. An intelligent camera system is developed to track lecturer and audience in [79] Panoramic cameras are used to capture the environment in [96, 53] The region of interests are detected from the panoramic scene and these parts are presented [96]. This approach is not applicable to previously developed videos. A spatially indexed camera is explained for navigation in [53] Our goal is to enable spatial browsing of recorded video. We do not assume that the scene is predefined. We consider scenes that are captured using a single camera. ....
[Article contains additional citation context not shown here]
Xinding Sun, Jonathan Foote, Don Kimber, and B.S. Manjunath. Panoramic video capturing and compressed domain virtual camera control. In ACM Multimedia 2001.
....and the other is inter coded. In a special case, when the macroblock perfectly matches its reference, it is skipped and not coded at all. To simplify the illustration, the skipped frames are categorized the same as No MC inter coded frames. The No MC inter coded macroblock has zero motion [8]. According to the definition of inter coded No MC, when the content of video changes are not too significant, and thus many macroblocks match their references, the inter coded No MC macroblock numbers would be high in a P frame. For example, pauses in sports games often coincide with small ....
. X. Sun, J. Foote, D. Kimber, B. S. Manjunath, "Panoramic Video Capturing and Compressed Domain Virtual Camera Control," Proc. ACM Multimedia, pp. 329-338, 2001.
....other wide angle cameras for seminar capturing [7] In general these systems involve camera motion, which is not helpful in the recognition process. Therefore, we choose the Flycam [3] panoramic system to capture the activities. For the system, we have designed virtual camera control methods [9][10] to output ROI video that covers the speaker. The advantage of this system is that it is automatic, and the speaker is always in the scene. There is no physical camera motion in the system and the virtual camera parameters are readily available for recognition purpose. Figure 1 gives one example ....
....speaker. The camera system is fixed and covers all the area where the speaker activities take place. The panoramic system produces real time panoramic video output. While we can compress the panoramic video first and extract ROI video from compressed domain for later activity recognition [10], here we choose to extract ROI video in real time [9] from the panoramic video. By doing this, we can avoid storing extra large amount of redundant (a) Panoramic view (b)ROI output Figure 1. Panoramic capturing and ROI output. data outside the ROI area in the panoramic video and still ....
.X. Sun, J. Foote, D. Kimber, B. S. Manjunath, "Panoramic Video Capturing and Compressed Domain Virtual Camera Control", Proc. ACM Multimedia, pp. 329-338, 2001.
No context found.
Sun, X., Foote, J., Kimber, D., and Manjunath, "Panoramic video capturing and compressed domain virtual camera control", ACM Multimedia, pp. 229-238, 2001.
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