Abstract: Existing remote workspace sharing systems are difficult to use. For example, changes made on a common work product by one user often appear abruptly on displays viewed by remote users. As a result the interaction is perceived as unnatural by the users and is often inefficient. Images of a display of a common work product are received from a camera at a first location. These images may also comprise information about objects between the display and the camera such as a user's hand editing a document on a tablet PC. These images are combined with images of the shared work product and displayed at remote locations. Advance information about remote user actions is then visible and facilitates collaborative mediation between users. Depth information may be used to influence the process of combining the images.

Abstract: There is a need to provide simple, accurate, fast and computationally inexpensive methods of object and hand pose recognition for many applications. For example, to enable a user to make use of his or her hands to drive an application either displayed on a tablet screen or projected onto a table top. There is also a need to be able to discriminate accurately between events when a user's hand or digit touches such a display from events when a user's hand or digit hovers just above that display. A random decision forest is trained to enable recognition of hand poses and objects and optionally also whether those hand poses are touching or not touching a display surface. The random decision forest uses image features such as appearance, shape and optionally stereo image features. In some cases, the training process is cost aware. The resulting recognition system is operable in real-time.

Abstract: Existing remote workspace sharing systems are difficult to use. For example, changes made on a common work product by one user often appear abruptly on displays viewed by remote users. As a result the interaction is perceived as unnatural by the users and is often inefficient. Images of a display of a common work product are received from a camera at a first location. These images may also comprise information about objects between the display and the camera such as a user's hand editing a document on a tablet PC. These images are combined with images of the shared work product and displayed at remote locations. Advance information about remote user actions is then visible and facilitates collaborative mediation between users. Depth information may be used to influence the process of combining the images.

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. Motion events in image sequences may be detected without explicit velocity computation.

Abstract: There is a need to provide simple, accurate, fast and computationally inexpensive methods of object and hand pose recognition for many applications. For example, to enable a user to make use of his or her hands to drive an application either displayed on a tablet screen or projected onto a table top. There is also a need to be able to discriminate accurately between events when a user's hand or digit touches such a display from events when a user's hand or digit hovers just above that display. A random decision forest is trained to enable recognition of hand poses and objects and optionally also whether those hand poses are touching or not touching a display surface. The random decision forest uses image features such as appearance, shape and optionally stereo image features. In some cases, the training process is cost aware. The resulting recognition system is operable in real-time.

Abstract: Foreground and background image segmentation is described. In an example, a seed region is selected in a foreground portion of an image, and a geodesic distance is calculated from each image element to the seed region. A subset of the image elements having a geodesic distance less than a threshold is determined, and this subset of image elements are labeled as foreground. In another example, an image element from an image showing at least a user, a foreground object in proximity to the user, and a background is applied to trained decision trees to obtain probabilities of the image element representing one of these items, and a corresponding classification assigned to the image element. This is repeated for each image element. Image elements classified as belonging to the user are labeled as foreground, and image elements classified as foreground objects or background are labeled as background.

Abstract: Image segmentation using star-convexity constraints is described. In an example, user input specifies positions of one or more star centers in a foreground to be segmented from a background of an image. In embodiments, an energy function is used to express the problem of segmenting the image and that energy function incorporates a star-convexity constraint which limits the number of possible solutions. For example, the star-convexity constraint may be that, for any point p inside the foreground, all points on a shortest path (which may be geodesic or Euclidean) between the nearest star center and p also lie inside the foreground. In some examples continuous star centers such as lines are used. In embodiments a user may iteratively edit the star centers by adding brush strokes to the image in order to progressively change the star-convexity constraints and obtain an accurate segmentation.

Abstract: Medical image rendering is described. In an embodiment a medical image visualization engine receives results from an organ recognition system which provide estimated organ centers, bounding boxes and organ classification labels for a given medical image. In examples the visualization engine uses the organ recognition system results to select appropriate transfer functions, bounding regions, clipping planes and camera locations in order to optimally view an organ. For example, a rendering engine uses the selections to render a two-dimensional image of medical diagnostic quality with minimal user input. In an embodiment a graphical user interface populates a list of organs detected in a medical image and a clinician is able to select one organ and immediately be presented with the optimal view of that organ. In an example opacity of background regions of the medical image may be adjusted to provide context for organs presented in a foreground region.

Abstract: Techniques for human body pose estimation are disclosed herein. Images such as depth images, silhouette images, or volumetric images may be generated and pixels or voxels of the images may be identified. The techniques may process the pixels or voxels to determine a probability that each pixel or voxel is associated with a segment of a body captured in the image or to determine a three-dimensional representation for each pixel or voxel that is associated with a location on a canonical body. These probabilities or three-dimensional representations may then be utilized along with the images to construct a posed model of the body captured in the image.

Abstract: Automatic identification of image features is described. In an embodiment, a device automatically identifies organs in a medical image using a decision forest formed of a plurality of distinct, trained decision trees. An image element from the image is applied to each of the trained decision trees to obtain a probability of the image element representing a predefined class of organ. The probabilities from each of the decision trees are aggregated and used to assign an organ classification to the image element. In another embodiment, a method of training a decision tree to identify features in an image is provided. For a selected node in the decision tree, a training image is analyzed at a plurality of locations offset from a selected image element, and one of the offsets is selected based on the results of the analysis and stored in association with the node.

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: Parallel processing for distance transforms is described. In an embodiment a raster scan algorithm is used to compute a distance transform such that each image element of a distance image is assigned a distance value. This distance value is a shortest distance from the image element to the seed region. In an embodiment two threads execute in parallel with a first thread carrying out a forward raster scan over the distance image and a second thread carrying out a backward raster scan over the image. In an example, a thread pauses when a cross-over condition is met until the other thread meets the condition after which both threads continue. In embodiments distances may be computed in Euclidean space or along geodesics defined on a surface. In an example, four threads execute two passes in parallel with each thread carrying out a raster scan over a different quarter of the image.

Abstract: A system and method for detecting and tracking targets including body parts and props is described. In one aspect, the disclosed technology acquires one or more depth images, generates one or more classification maps associated with one or more body parts and one or more props, tracks the one or more body parts using a skeletal tracking system, tracks the one or more props using a prop tracking system, and reports metrics regarding the one or more body parts and the one or more props. In some embodiments, feedback may occur between the skeletal tracking system and the prop tracking system.

Abstract: Given an image of structured and/or unstructured objects, semantically meaningful areas are automatically partitioned from the image, each area labeled with a specific object class. Shape filters are used to enable capturing of some or all of the shape, texture, and/or appearance context information. A shape filter comprises one or more regions of arbitrary shape, size, and/or position within a bounding area of an image, paired with a specified texton. A texton comprises information describing the texture of a patch of surface of an object. In a training process a sub-set of possible shape filters is selected and incorporated into a conditional random field model of object classes. The conditional random field model is then used for object detection and recognition.

Abstract: Methods and a processing device are provided for restoring pixels damaged by artifacts caused by dust, or other particles, entering a digital image capturing device. A user interface may be provided for a user to indicate an approximate location of an artifact appearing in a digital image. Dust attenuation may be estimated and an inverse transformation, based on the estimated dust attenuation, may be applied to damaged pixels in order to recover an estimate of the underlying digital image. One or many candidate source patch may be selected based on having smallest pixel distances, with respect to a target patch area. The damaged pixels included in the target patch area may be considered when calculating the pixel distance with respect to candidate source patches. RGB values of corresponding pixels of source patches may be used to restore the damaged pixels included in the target patch area.

Abstract: Given an image of structured and/or unstructured objects we automatically partition it into semantically meaningful areas each labeled with a specific object class. We use a novel type of feature which we refer to as a shape filter. Shape filters enable us to capture some or all of shape, texture and appearance context information. A shape filter comprises one or more regions of arbitrary shape, size and position within a bounding area of an image, paired with a specified texton. A texton comprises information describing the texture of a patch of surface of an object. In a training process we select a sub-set of possible shape filters and incorporate those into a conditional random field model of object classes. That model is then used for object detection and recognition.

Abstract: Image processing using geodesic forests is described. In an example, a geodesic forest engine determines geodesic shortest-path distances between each image element and a seed region specified in the image in order to form a geodesic forest data structure. The geodesic distances take into account gradients in the image of a given image modality such as intensity, color, or other modality. In some embodiments, a 1D processing engine carries out 1D processing along the branches of trees in the geodesic forest data structure to form a processed image. For example, effects such as ink painting, edge-aware texture flattening, contrast-aware image editing, forming animations using geodesic forests and other effects are achieved using the geodesic forest data structure. In some embodiments the geodesic forest engine uses a four-part raster scan process to achieve real-time processing speeds and parallelization is possible in many of the embodiments.

Abstract: A multi-layer graph for dense stereo dynamic programming can improve synthesis of cyclopean virtual images by distinguishing between stereo disparities caused by occlusion and disparities caused by non-fronto-parallel surfaces. This distinction can be leveraged to reduce image artifacts, such as “halos”. Distinguishing at least between these two types of disparities allows improved matching of left and right pixel data, which increases the amount of correct pixel information used in constructing the cyclopean virtual image and minimizes occlusion artifacts.

Abstract: A system for multi-modal mapping of images is described. Embodiments are described where the image mapping system is used for visualizing high dynamic range images such as medical images, satellite images, high dynamic range photographs and the like and also for compressing such images. In examples, high bit-depth images are tone-mapped for display on equipment of lower bit-depth without loss of detail. In embodiments, the image mapping system computes statistics describing an input image and fits a multi-modal model to those statistics efficiently. In embodiments, the multi-modal model is a Gaussian mixture model and a plurality of sigmoid functions corresponding to the multi-modal model are obtained. In an embodiment the sigmoid functions are added to form a tone-mapping function which is used to transform a high bit-depth image such as 16 or 12 bits per pixel to a low bit-depth image such as 8 bits per pixel.

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.