Abstract: The system and method of the present invention automatically assigns “scores” to the predictor/variable value pairs of a conventional probabilistic model to measure the relative impact or influence of particular elements of a set of topics, items, products, etc. in making specific predictions using the probabilistic model. In particular, these scores measure the relative impact, either positive or negative, that the value of each individual predictor variable has on the posterior distribution of the target topic, item, product, etc., for which a probability is being determined. These scores are useful for understanding why each prediction is make, and how much impact each predictor has on the prediction. Consequently, such scores are useful for explaining why a particular prediction or recommendation was made.

Abstract: The present invention involves a new system and process for graphically representing time dependent information or data objects, from at least one separate program, in a sequential manner using a dynamic time strip having implicit or explicit indications of time. In particular, the dynamic time strip is a linear graphical strip comprising graphical representations of time dependent information elements or data objects which may or may not include text or markers to indicate the time (i.e. second, minute, hour, day and/or date) along the length of the strip. As time passes, and data objects are changed, added or removed from the time strip, either automatically or through user interaction, the time strip is automatically and dynamically updated to reflect these changes along with showing the passage of time. The period of time represented by the dynamic time strip is preferably one day, with the dynamic time strip automatically scrolling forward as time passes.

Abstract: The system and method of the present invention automatically provides dynamically generated completion information for facilitating user input of email addresses or contact information. This completion information is developed from a “data store” comprised of multiple data sources such as previously sent or received email, and other types of electronic files such as word processor or spreadsheet files. The present invention monitors and uses the information in the data store to automatically store, track, maintain, and organize data entries in a dynamic “resolution list”. As a user begins to input an email address or contact, the present invention can either automatically complete the entry using a most probable result from the resolution list, or can display a list of likely matches from which the user may select the desired email address or contact.

Abstract: The present invention is embodied in a system and method for performing sparse transformed template matching using three dimensional rasterization by comparing and matching a first set of digital data to at least a second set of digital data. The system and method includes raster transforming at least one of the first set of digital data and the second set of digital data, placing multiple images of the digital data in texture memory as multiple textures, gathering statistics between the textures and statistically comparing and matching the raster transformed sets of digital data to appropriately corresponding portions of each other. The first or the second set of digital data can be transformed during statistical analysis to enhance statistical analysis of the digital data.

Abstract: A system and process for streaming delivery of dynamically scalable media content over a network, such as, for example, the Internet or a wireless network, while automatically accounting for both fluctuating network bandwidth and packet loss. A system of rate-distortion based packet selection and organization is used to maximize the quality of streamed media files that have been encoded using any conventional scalable encoder. Media file quality is maximized for available bandwidth by scoring packets comprising encoded media files based on their contribution to the quality of a reconstructed media file, then preferentially transmitting those packets having the highest scores. In addition, where packets are lost during transmission, those packets that will provide the maximum rate-distortion decrease, are preferentially retransmitted prior to other lost packets which, if transmitted in the same time slot, would provide a lesser rate-distortion decrease.

Abstract: The present invention is embodied in a system and process for identifying and locating people and objects of interest in a scene by selectively clustering distinct three-dimensional regions or “blobs” within the scene and comparing the blob clusters to a model. Specifically, a background subtraction process is used to generate a working image from a baseline depth image and a live depth image of the scene. Distinct blobs in the working image are selectively clustered to generate “candidate blob clusters” which are then compared to a model representing the people or objects of interest. The comparison of candidate blob clusters to the model identifies the blob clusters that most accurately represent the people or objects of interest in the scene by determining the closest match or matches to the model. Blob clusters may be compared to a plurality of models representing people or objects of different sizes and shapes.

Abstract: An “object extractor” automatically identifies and segments repeating media objects in a media stream. “Objects” are any section of non-negligible duration, i.e., a song, video, advertisement, jingle, etc., which would be considered to be a logical unit by a human listener or viewer. Identification and segmentation of repeating objects is achieved by directly comparing sections of the media stream to identify matching portions of the stream, then aligning the matching portions to identify object endpoints. Alternately, a suite of object dependent algorithms is employed to target particular aspects of the stream for identifying possible objects within the stream. Confirmation of possible objects as repeating objects is achieved by automatically searching for potentially matching objects in a dynamic object database, followed by a detailed comparison to one or more of the potentially matching objects.

Abstract: The present invention involves a new system and method for synthesizing textures from an input sample. A system and method according to the present invention uses a unique accelerated patch-based sampling system to synthesize high-quality textures in real-time using a small input texture sample. The patch-based sampling system of the present invention works well for a wide variety of textures ranging from regular to stochastic. Potential feature mismatches across patch boundaries are avoided by sampling patches according to a non-parametric estimation of the local conditional Markov Random Field (MRF) density function.

Abstract: The present invention is embodied in a system and process for automatically learning a reliable tracking system. The tracking system is learned by using information produced by an initial object model in combination with an initial tracking function, and a data acquisition function for gathering observations about each image. The initial tracking function probabilistically determines the configuration of one or more target objects in a temporal sequence of images. The observations gathered by the data acquisition function include information that is relevant to parameters desired for a final object model. These relevant observations may include information such as the color, shape, or size of a tracked object, and depend on the parameters necessary to support the final tracking function.

Abstract: The present invention involves a new system and process for image-based relativistic rendering of one or more single or sequential images of a scene. Such images may be produced by any conventional method, and include, for example, photographs, panoramas, motion video or film, or any other type of image captured by a camera or other image capture device. Relativistic rendering in accordance with the present invention transforms one or more initial images to produce photo-realistic images that accurately simulate how the scene would be viewed by an observer that is traveling at any desired velocity up to the speed of light. These rendered images are equal in resolution and quality to the initial images. Such relativistic simulation is based on Einstein's Special Theory of Relativity. Further, unlike previous techniques, relativistic rendering in accordance with the present invention does not depend upon artificial or synthetic models, or upon ray-tracing methods.

Abstract: A minimum sampling rate and a minimum sampling curve for continuous representation of a scene are disclosed. The minimum sampling rate for light-field rendering is determined in accordance with Δ ⁢   ⁢ t max = 1 2 ⁢ K Ω v ⁢   ⁢ fh d , where K&OHgr;v accounts for a light-field signal cut-off frequency, a sampling camera resolution and an output resolution, ƒ specifies a sampling camera focal length, and hd specifies a light-field depth range.

Abstract: A training system for a classifier utilizes both a back-propagation system to iteratively modify parameters of functions which provide raw output indications of desired categories, wherein the parameters are modified based on a weighted decay, and a probability determining system with further parameters that are determined during iterative training. A margin error metric may be combined with weight decay, and a sigmoid is used to calibrate the raw outputs to probability percentages for each category. A method of training such a system involves gathering a training set of inputs and desired corresponding outputs. Classifier parameters are then initialized and an error margin is calculated with respect to the classifier parameters. A weight decay is then used to adjust the parameters.

Abstract: Automated layer extraction from 2D images making up a 3D scene, and automated image pixel assignment to layers, to provide for scene modeling, is disclosed. In one embodiment, a computer-implemented method determines a number of planes, or layers, and assigns pixels to the planes. The method can determine the number of planes by first determining the high-entropy pixels of the images, and then determining a 1-plane through a predetermined n-plane estimation, such as via a robust estimation, and a most likely x-plane estimation, where x is between 1 and n, such as via a Bayesian approach. Furthermore, the method can assign pixels via an iterative EM approach based on classifying criteria.

Abstract: A system and method for improving the uniformity in exposure and tone of a digital image using a locally adapted histogram equalization approach. This approach involves first segmenting the digital image into a plurality of image patches. For each of these patches, a pixel brightness level histogram is created. The histogram for each patch is then optionally averaged with the histograms associated with a prescribed number of neighboring image patches. A normalized cumulative distribution function is generated for each patch based on the associated averaged histogram. This normalized-cumulative distribution function identifies a respective new pixel brightness level for each of the original pixel brightness levels. For each of the original pixel brightness levels, the 1s associated new pixel brightness levels from one or more of the image patches are blended. Preferably, this blending is accomplished using either a bilinear or biquadratic interpolator function.

Abstract: A disposable container according to present invention is embodied in a thin-walled container having an integral sealing flap. Such containers include cups or other vessels for holding beverages, food or other substances. The disposable container preferably has a truncated conical shape. One example of such a shape is a typical beverage cup commonly dispensed at fast food restaurants and convenience stores. However, any shape or size of disposable container may be used, so long as the container is sealed using an integral sealing flap in accordance with the present invention. Once the container has been filled it is closed by squeezing or folding the top of the container together. The container is then sealed by folding the integral sealing flap over one side of the top of the closed container. In one embodiment, the sealing flap is then secured using conventional adhesives that are integral to either or both the surface of the container, and the surface of the sealing flap.

Abstract: A disposable container according to present invention is embodied in a thin-walled container having an integral sealing flap. Such containers include cups or other vessels for holding beverages, food or other substances. The disposable container preferably has a truncated conical shape. One example of such a shape is a typical beverage cup commonly dispensed at fast food restaurants and convenience stores. However, any shape or size of disposable container may be used, so long as the container is sealed using an integral sealing flap in accordance with the present invention. Once the container has been filled, it is closed by squeezing or folding the top of the container together. The container is then sealed by folding the integral sealing flap over one side of the top of the closed container. The sealing flap is then secured using conventional adhesives, such as, for example, glue, rubber cement, or tape, that are integral to either or both the surface of the container, or the surface of the sealing flap.

Abstract: Correction of errors and losses in a receiver-driven layered multicast (RLM) of real-time media over a network is augmented using one or more layers of error correction information. Each receiver separately optimizes the quality of received information by subscribing to at least one error correction layer. Ideally, each source layer in a RLM has one or more associated multicasted error correction data streams. Each error correction layer contains information for replacing lost packets from the associated source layer. More than one error correction layer is proposed to correct for lost packets in other error correction layers. Error correction streams are preferably generated using a pseudo-Automatic Repeat Request (ARQ) wherein a broadcaster sends both the source packets in a primary stream and delayed versions thereof in one or more redundant streams. A hybrid technique combines the psuedo-ARQ method with an adaptation of Forward Error Correction (FEC) techniques.

Abstract: A container according to the present invention involves a thin lightweight case, or receptacle of any desired size or shape for holding or storing objects such as, eyeglasses, handheld electronic devices, pagers, cell phones, etc. The container is manufactured from any conventional material such as, a cloth, fabric, soft leather, rubber, plastic, or polymer material, or some combination thereof, and is preferably thin enough to be easily inserted into a user's pocket. The outer surface of the container is designed to provide a friction force when in a user's pocket, or in contact with a surface such as a car dash or a boat deck, for preventing the container from slipping out of the user's pocket, or from sliding along or falling off of the surface. Additionally, in one embodiment, the container is configured to prevent or reduce shock and vibration from being transmitted to objects within the container.

Abstract: “Correction of errors and losses occurring during a receiver-driven layered multicast (RLM) of real-time media over a heterogeneous packet network such as the Internet is accomplished by augmenting RLM with one or more layers of error correction information. Each receiver separately optimizes the quality of received audio and video information by subscribing to at least one error correction layer. Ideally, each source layer in a RLM would have one or more associated multicasted error correction data streams (i.e., layers). Each error correction layer contains information that can be used to replace lost packets from the associated source layer. More than one error correction layer is proposed as some of the error correction packets contained in the data stream needed to replace the packets lost in the associated source stream may themselves be lost in transmission.