Abstract: A system is provided for selecting a foreground region of an image, given a set of pixels defining the boundary of the foreground region of the image. The system includes a component to determine a foreground value (F) and opacity value (?) for each pixel on the set wherein ?is determined via a subpixel edge-offset to facilitate a separation the foreground region from background portions of the image. The foreground value (F) and opacity value (?) are then employed to smoothly mix the foreground region of the image with a subsequent image background region.

Abstract: Artifacts are detected in a cyclopean virtual image generated from stereo images. A disparity map is generated from the stereo images. Individual projected images are determined based on the disparity map and the corresponding stereo images. A difference map is computed between the individual projected images to indicate the artifacts. A source patch in the virtual image is defined relative to an artifact. A replacement target patch is generated using a split-patch search technique as a composite of a background exemplar patch and a foreground exemplar patch. Each exemplar patch may be identified from an image patch selected from at least one of the stereo images. The source patch of the virtual image is replaced by the replacement target patch to correct the detected artifact.

Abstract: Systems and methods for object or pattern detection that use a nonlinear support vector (SV) machine are described. In the illustrated and described embodiment, objects or patterns comprising faces are detected. The decision surface is approximated in terms of a reduced set of expansion vectors. In order to determine the presence of a face, the kernelized inner product of the expansion vectors with the input pattern are sequentially evaluated and summed, such that if at any point the pattern can be rejected as not comprising a face, no more expansion vectors are used. The sequential application of the expansion vectors produces a substantial saving in computational time.

Abstract: A real-time stereo video signal of a captured scene with a physical foreground object and a physical background is received. In real-time, a foreground/background separation algorithm is used on the real-time stereo video signal to identify pixels from the stereo video signal that represent the physical foreground object. A video sequence is produced by rendering a 3d virtual reality based on the identified pixels of the physical foreground object.

Abstract: A low spatial frequency video sequence is enhanced to provide a higher spatial frequency video sequence using a super-resolution process. Patches of higher frequency image data are inferred from the images of the lower frequency video sequence and a dictionary containing a training set of higher resolution image data. An inference module selects result patches from the training set to preserve spatial consistency within each image frame and to preserve temporal consistency, between image frames of the resulting video sequence. Temporal consistency can be preserved for static portions and/or moving portions of each frame.

Abstract: It is required to provide an automated process for forming a visually appealing digital collage (also referred to as a tapestry or photomontage) from a plurality of input images. Input images and a collage are labeled. Labels in the collage specify which regions of the input images are used to form those collage regions. An energy function is created for the labeling which contains terms that we design to enable us to take into account various criteria we specify to produce a good collage. For example, those criteria may relate to the amount of information in the images, similarity of the images, entropy in the images or object class information. A multi-stage optimization process is applied to the energy function which is fast and enables large numbers of input images to be used to produce a collage.

Abstract: A method, apparatus, and article of manufacture altering a displayed image presented to a user on a viewing device using a user-controlled orientation of the viewing device to determine how the displayed image is to be presented. The viewing device includes a plurality of tilt sensors that are used to determine the orientation of the viewing device. As the user moves the orientation of the viewing device, the tilt sensors detect the change in the device orientation. These changes in orientation are used to alter the image being displayed upon the viewing device.

Abstract: Real-time segmentation of foreground from background layers in binocular video sequences may be provided by a segmentation process which may be based on one or more factors including likelihoods for stereo-matching, color, and optionally contrast, which may be fused to infer foreground and/or background layers accurately and efficiently. In one example, the stereo image may be segmented into foreground, background, and/or occluded regions using stereo disparities. The stereo-match likelihood may be fused with a contrast sensitive color model that is initialized or learned from training data. Segmentation may then be solved by an optimization algorithm such as dynamic programming or graph cut. In a second example, the stereo-match likelihood may be marginalized over foreground and background hypotheses, and fused with a contrast-sensitive color model that is initialized or learned from training data. Segmentation may then be solved by an optimization algorithm such as a binary graph cut.

Abstract: The present invention involves a new system and method for probabilistic exemplar-based tracking of patterns or objects. Tracking is accomplished by first extracting a set of exemplars from training data. The exemplars are then clustered using conventional statistical techniques. Such clustering techniques include k-medoids clustering which is based on a distance function for determining the distance or similarity between the exemplars. A dimensionality for each exemplar cluster is then estimated and used for generating a probabilistic likelihood function for each exemplar cluster. Any of a number of conventional tracking algorithms is then used in combination with the exemplars and the probabilistic likelihood functions for tracking patterns or objects in a sequence of images, or in a space, or frequency domain.

Abstract: Segmentation of foreground from background layers in an image may be provided by a segmentation process which may be based on one or more factors including motion, color, contrast, and the like. Color, motion, and optionally contrast information may be probabilistically fused to infer foreground and/or background layers accurately and efficiently. A likelihood of motion vs. non-motion may be automatically learned from training data and then fused with a contrast-sensitive color model. Segmentation may then be solved efficiently by an optimization algorithm such as a graph cut.

Abstract: Images of the same scene from multiple cameras may be use to generate a stereo disparity map. At least a portion of the stereo disparity map may be compared to a kernel image to detect and/or determine the location of an object in the disparity map. The kernel image is an array of pixel values which represent the stereo disparity of an object to be located, more particularly, the kernel image indicates the 3-dimensional surface shape of the object to be located from a point of view. The disparity map containing the located object may be process to manipulate the display of the stereo-based image and/or an input image.

Abstract: Systems and methods for object or pattern detection that use a nonlinear support vector (SV) machine are described. In the illustrated and described embodiment, objects or patterns comprising faces are detected. The decision surface is approximated in terms of a reduced set of expansion vectors. In order to determine the presence of a face, the kernelized inner product of the expansion vectors with the input pattern are sequentially evaluated and summed, such that if at any point the pattern can be rejected as not comprising a face, no more expansion vectors are used. The sequential application of the expansion vectors produces a substantial saving in computational time.

Abstract: An example-based filling system identifies appropriate filling material to replace a destination region in an image and fills the destination region using this material, thereby alleviating or minimizing the amount of manual editing required to fill a destination region in image. Tiles of image data are borrowed from the proximity of the destination region or some other source to generate new image data to fill in the region. Destination regions may be designated by user input (e.g., selection of an image region by a user) or by other means (e.g., specification of a color or feature to be replaced).

Abstract: A portable anchor for use on a roof (26), the anchor comprising a force distribution member (12) and connection means (13) for receiving a load support line (28) wherein the connection means is positioned around the junction between the attachment member and the force distribution member (12). The attachment member (11) is preferably an elongate arrangement with a gap between a first side structure and a second side structure in the form of a clamping arrangement. The gap is dimensioned to receive an edge (25) of the roof and preferably has a stop member (15) positionable against an edge of the roof cladding. The force distribution member (12) is preferably dimensioned and configured to mate with corrugations (27) or other indentations in the roof to resist lateral displacement. The connection means (13) is positioned towards the junction of the arm and the longate member to resist the tendency for lateral rotation and displacement of the roof anchor as a harnessed worker moves laterally on the roof.

Abstract: An output image formed from at least a portion of one or more input images may be automatically synthesized as a tapestry image. To determine which portion or region of each input image will be used in the image tapestry, the regions of each image may be labeled by one of a plurality of labels. The multi-class labeling problem of creating the tapestry may be resolved such that each region in the tapestry is constructed from one or more salient input image regions that are selected and placed such that neighboring blocks in the tapestry satisfy spatial compatibility. This solution may be formulated using a Markov Random Field and the resulting tapestry energy function may be optimized in any suitable manner. To optimize the tapestry energy function, an expansion move algorithm for energy functions may be generated to apply to non-metric hard and/or soft constraints.

Abstract: The present invention involves a new system and method for probabilistic exemplar-based tracking of patterns or objects. Tracking is accomplished by first extracting a set of exemplars from training data. The exemplars are then clustered using conventional statistical techniques. Such clustering techniques include k-medoids clustering which is based on a distance function for determining the distance or similarity between the exemplars. A dimensionality for each exemplar cluster is then estimated and used for generating a probabilistic likelihood function for each exemplar cluster. Any of a number of conventional tracking algorithms is then used in combination with the exemplars and the probabilistic likelihood functions for tracking patterns or objects in a sequence of images, or in a space, or frequency domain.

Abstract: The present invention involves a new system and method for probabilistic exemplar-based tracking of patterns or objects. Tracking is accomplished by first extracting a set of exemplars from training data. The exemplars are then clustered using conventional statistical techniques. Such clustering techniques include k-medoids clustering which is based on a distance function for determining the distance or similarity between the exemplars. A dimensionality for each exemplar cluster is then estimated and used for generating a probabilistic likelihood function for each exemplar cluster. Any of a number of conventional tracking algorithms is then used in combination with the exemplars and the probabilistic likelihood functions for tracking patterns or objects in a sequence of images, or in a space, or frequency domain.

Abstract: Techniques are disclosed to provide more efficient and improved border matting for extracted foreground images, e.g., without requiring excessive user interaction. Border matting techniques described herein generate relatively continuous transparency (or alpha values) along the boundary of the extracted object (e.g., limiting color bleeding and/or artifacts).

Abstract: An example-based filling system identifies appropriate filling material to replace a destination region in an image and fills the destination region using this material, thereby alleviating or minimizing the amount of manual editing required to fill a destination region in image. Tiles of image data are borrowed from the proximity of the destination region or some other source to generate new image data to fill in the region. Destination regions may be designated by user input (e.g., selection of an image region by a user) or by other means (e.g., specification of a color or feature to be replaced). In addition, the order in which the destination region is filled by example tiles may be configured to emphasize the continuity of linear structures and composite textures using a type of isophote-driven image-sampling process.

Abstract: Techniques are disclosed to provide more efficient and improved extraction of a portion of a scene without requiring excessive user interaction. More particularly, the extraction may be achieved by using iterated graph cuts. In an implementation, a method includes segmenting an image into a foreground portion and a background portion (e.g., where an object or desired portion to be extracted is present in the foreground portion). The method determines the properties corresponding to the foreground and background portions of the image. Distributions may be utilized to model the foreground and background properties. The properties may be color in one implementation and the distributions may be a Gaussian Mixture Model in another implementation. The foreground and background properties are updated based on the portions. And, the foreground and background portions are updated based on the updated foreground and background properties.