Abstract: Object tracking includes an audio model that receives at least two audio input signals and a video model that receives a video input. The audio model and the video model employ probabilistic generative models which are combined to facilitate object tracking. Expectation maximization can be employed to modify trainable parameters of the audio model and the video model.

Abstract: A stereoscopic image pickup apparatus includes a first imaging section, a second imaging section, a zoom controller, and an image selection section. The first imaging section includes a zoom lens. The second imaging section includes a zoom lens. The zoom controller controls angles of view of the zoom lenses of the first imaging section and the second imaging section. The image selection section outputs image signals, which are output by the first imaging section and the second imaging section, as image signals of two channels constituting a stereoscopic image when the angles of view controlled by the zoom controller are equal to or greater than a predetermined value. In addition, the image selection section outputs image signals, which are based on the image signal output by either the first imaging section or the second imaging section, as the image signals of two channels constituting the stereoscopic image when the angles of view controlled by the zoom controller are less than the predetermined value.

Abstract: A device and a method for non-contact recording of spatial coordinates of a surface includes a sensor head and a controller connected to the sensor head. The sensor head is designed as a portable hand-held device, a projector and two cameras. The controller is also designed as a portable device, with an electrochemical energy source housed in the controller or in the sensor head for power supply to the sensor head and the controller and no cable connection of the sensor head or the controller is provided to further stationary components of the device.

Abstract: Automated or semi-automated conversion of 2D movies into 3D movies through generation and use of director-style statistics or characteristics to determine which processes to apply to convert regions of images based on the director's movie making style, without requiring the creation, moving or reshaping of masks. System works by breaking a movie down into scenes and shots and identifying lighting, lens, camera moves and color schemes habitually utilized by a director. The characteristics or statistical information is stored in a database accessible by a computer over a network for example. Swarms of computers or any other architecture employing any required amount of computing power allows for the desired speed of conversion to take place. Once a director's style has been analyzed, embodiments utilize the director-style characteristics to determine the specific processes to utilize to convert the movie from 2D to 3D.