| 摘要 |
A method comprises: combining 210 visual information 202 from an image of at least an object feature, and information 209 from a sensor attached to the object; and determining the position of a representation of the feature in a virtual representation of the object. As also independently claimed, the capture images include content representing plural markers attached to a motion capture body suit 102, with the physical position sensor data provided by plural inertial sensors, attached to an actor's body. The method and system may determine regions of future images likely to include a feature, based on sensor data, allowing motion tracking. The object with the physical (e.g. inertial) sensor may itself be a camera. The virtual representation may be a virtual online game character, and the visual information may be an object silhouette. Sensors may provide time series data. One or more poses (e.g. position and orientation) of the object can be represented. |
