摘要:
Estimating a pose of an articulated 3D object model (4) by a computer is done by •obtaining a sequence of source images (10) and therefrom corresponding source image segments (13) with objects (14) separated from the image background; •matching such a sequence (51) with sequences (52) of reference silhouettes (13′), determining one or more selected sequences of reference silhouettes (13′) forming a best match; •for each of these selected sequences of reference silhouettes (13′), retrieving a reference pose that is associated with one of the reference silhouettes (13′); and •computing an estimate of the pose of the articulated object model (4) from the retrieved reference pose or poses. The result of these steps is an initial pose estimate, which then can be used in further steps, for example, for maintaining local consistency between pose estimates from consecutive frames, and global consistency over a longer sequence of frames.
摘要:
Estimating a pose of an articulated 3D object model (4) by a computer is done by •obtaining a sequence of source images (10) and therefrom corresponding source image segments (13) with objects (14) separated from the image background; •matching such a sequence (51) with sequences (52) of reference silhouettes (13′), determining one or more selected sequences of reference silhouettes (13′) forming a best match; •for each of these selected sequences of reference silhouettes (13′), retrieving a reference pose that is associated with one of the reference silhouettes (13′); and •computing an estimate of the pose of the articulated object model (4) from the retrieved reference pose or poses. The result of these steps is an initial pose estimate, which then can be used in further steps, for example, for maintaining local consistency between pose estimates from consecutive frames, and global consistency over a longer sequence of frames.
摘要:
Using photographic flash for candid shots often results in an unevenly lit scene, in which objects in the back appear dark. A spatially adaptive photographic flash (100) is disclosed, in which the intensity of illumination (21, 23) varies depending on the depth and reflectivity (30, 101) of features in the scene. Adaption to changes in depth are used in a single-shot method. Adaption to changes in reflectivity are used in a multishot method. The single-shot method requires only a depth image (30), whereas the multi-shot method requires at least one color image (40) in addition to the depth data (30).
摘要:
Using photographic flash for candid shots often results in an unevenly lit scene, in which objects in the back appear dark. A spatially adaptive photographic flash (100) is disclosed, in which the intensity of illumination (21, 23) varies depending on the depth and reflectivity (30, 101) of features in the scene. Adaption to changes in depth are used in a single-shot method. Adaption to changes in reflectivity are used in a multishot method. The single-shot method requires only a depth image (30), whereas the multi-shot method requires at least one color image (40) in addition to the depth data (30).
摘要:
Provided is point-based efficient three-dimensional (3D) information representation from a color image that is obtained from a general Charge-Coupled Device (CCD)/Complementary Metal Oxide Semiconductor (CMOS) camera, and a depth image that is obtained from a depth camera. A 3D image processing method includes storing a depth image associated with an object as first data of a 3D data format, and storing a color image associated with the object as color image data of a 2D image format, independent of the first data.
摘要:
An image display apparatus includes a display layer and an image separation layer. The image separation layer is adapted to separate a displayed image on the display layer into a first image for a left eye of an observer and into a second image for a right eye of an observer. This renders the image display apparatus autostereoscopic. In order to improve the appearance of the displayed image, the apparatus is adapted to activate the display layer and the image separation layer substantially only within a given two-dimensional silhouette and to remain substantially transparent outside of the silhouette. In other aspects, the apparatus may be adapted to dynamically adjust an observation angle between the first image and the second image by controlling the distance between the display layer and the image separation layer by an actuator. The apparatus may be mounted on a robotic unit.
摘要:
Systems, methods and articles of manufacture are disclosed for stereoscopically editing video content. In one embodiment, image pairs of a sequence may be stereoscopically modified by altering at least one image of the image pair. The at least one image may be altered using at least one mapping function. The at least one image may also be altered based on a saliency of the image pair. The at least one image may also be altered based on disparities between the image pair. Advantageously, stereoscopic properties of video content may be edited more conveniently and efficiently.
摘要:
Techniques are disclosed for performing image space reprojection iteratively. An insignificant parallax threshold depth is computed for a source image. Portions of the image having depth values greater than the insignificant parallax threshold depth may be shifted uniformly to produce corresponding portions of the reprojection (target) image. An iterative fixed-point reprojection algorithm is used to reproject the portions of the source image having depth values less than or equal to the insignificant parallax threshold depth. The fixed point reprojection algorithm quickly converges on the best pixel in the source image for each pixel in a target image representing an offset view of the source image. An additional rendering pass is employed to fill disoccluded regions of the target image, where the reprojection algorithm fails to converge.
摘要:
Systems, methods and articles of manufacture are disclosed for performing scalable video coding. In one embodiment, non-linear functions are used to predict source video data using retargeted video data. Differences may be determined between the predicted video data and the source video data. The retargeted video data, the non-linear functions, and the differences may be jointly encoded into a scalable bitstream. The scalable bitstream may be transmitted and selectively decoded to produce output video for one of a plurality of predefined target platforms.