Abstract: Various embodiments of the present invention are directed to object segmentation of digital video streams and digital images. In one aspect, a method for segmenting an object in a digitally-encoded image includes designing a non-linear local function that generates function values associated with identifying the object of interest and a combination function that combines the function values, the object of interest encoded in digital image data. The method includes forming orthogonal projections of the digital image data based on the function values and the combination function. In addition, the orthogonal projections can be used to determining boundaries segmenting the object in the digital image data. The method also includes extracting an image of the object that lies within the boundaries.

Abstract: An input image (7) having a first pixel resolution is acquired from an image capture system (2). A respective characterization of each of at least one visual quality feature of the input image (7) is determined. An output thumbnail image (9) is produced from the input image (7). The output thumbnail image (9) reflects the respective characterization of each visual quality feature. The output thumbnail image (9) has a second pixel resolution lower than the first pixel resolution. The output thumbnail image (9) is output in association with operation of the image capture system (2).

Abstract: In a method of processing an input image, compression parameters of the input image are determined and luminance artifacts in the input image are reduced to obtain an improved luminance component of the input image, where the determined compression parameters guide the luminance artifact reduction. In addition, a chrominance noise and chrominance artifact level of the input image is estimated and a spatially-correlated chrominance noise and chrominance artifact of the input image is reduced with the improved luminance component, where the estimated chrominance noise and chrominance artifact level guide the spatially-correlated chrominance noise and chrominance artifact reduction to obtain an improved chrominance component of the input image. Moreover, an output image having the improved luminance component and the improved chrominance component is outputted.

Abstract: A system for generating a media playlist comprising a media management module operable to select a first media item from a plurality of media items stored in a media database for playback; and using raw user input data representing a measure of the popularity of the first media item, generate preference data representing a refined user preference for the first media item; wherein the preference data is used to determine a second media item from the plurality of media items for playback.

Abstract: Systems and computer readable mediums storing computer executable programs for generating a media playlist are disclosed. A first media item is selected from a plurality of media items for playback. A first determination is made regarding the number of times each of the plurality of media items was accessed for playback following an access for playback of the first media item. Each of the plurality of media items is weighted such that the probability of stochastically selecting each of the plurality of media items as a second media item for playback following a playback of the first media item is based on the first determination. The second media item is stochastically selected from the weighted plurality of media items for playback following the first media item.

Abstract: Various aspects of the present invention are directed to systems and methods for highlighting participant activities in video conferencing. In one aspect, a method of generating a dynamic visual representation of participants taking part in a video conference comprises rendering an audio-visual representation of the one or more participants at each site taking part in the video conference using a computing device. The method includes receiving a saliency signal using the computing device, the saliency signal identifying the degree of current and/or recent activity of the one or more participants at each site. Based on the saliency signal associated with each site, the method applies image processing to elicit visual popout of active participants associated each site, while maintaining fixed scale and borders interface of the visual representation of the one or more participants at each site.

Abstract: A method and system are disclosed for capturing media during a recording session using a separate input device for each of plural audio sources. An exemplary method includes producing audio data with the separate input device, processing the audio data to identify a portion of audio data having a first audio characteristic, and storing an audio record for each identified portion of audio data. Each audio record is associated with temporal data used in determining a sequence of the identified portion of audio data in relation to other identified portions of audio data from other separate input devices, and each audio record is associated with identity data representing identifying characteristics for the identified portion of audio data.

Abstract: The present invention describes a visual-collaborative system comprising: a display screen having a first surface and a second surface; a first projector positioned to project images onto a projection surface of the display screen, wherein the projected images can be observed by viewing the second surface; and a first camera system positioned to capture images of objects through the display screen, the first camera system including a first filter disposed between a first camera and the first surface, wherein the first filter passes the light received by the camera but substantially blocks the light produced by the first projector, wherein the first filter is a GMR (Guided Mode Resonance) filter.

Abstract: For each of multiple image pixels, input color component values of the pixel in an input device-dependent color space are transformed to output color component values in an output device-dependent color space characterized by an output color gamut defined by a respective gamut range for each of the output color components. In this process, the input color component values of the pixel are multiplied with corresponding elements of a device-dependent characterization matrix to produce a set of product values. The output color component values are derived from the product values. The values of a particular one of the output color components are ascertained based on a continuous nonlinear companding function that maps a function input value derived from one or more of the product values to a function output value that increases monotonically with increasing function input values over the respective gamut range of the particular output color component.

Abstract: Embodiments of the present invention are directed to a visual-collaborative system enabling geographically distributed groups to engage in face-to-face, interactive collaborative video conferences. In one aspect, a visual-collaborative system comprises a display screen, a camera system, and a projector. The display screen has a first surface and a second surface, and the camera system is positioned to capture images of objects through the display screen. The projector is positioned to project images onto a projection surface of the display screen, wherein the projected images can be observed by viewing the second surface. The system includes a first filter disposed between the camera and the first surface, where the first filter passes light receives by the camera but substantially blocks the light that is produced by the projector.

Abstract: Systems and methods of interfacing with a machine are described. In one aspect, sets of contemporaneous images of an interactive space are acquired from multiple respective fields of view. An input target is detected in the acquired images. Coordinates of the input target detected in the acquired images are computed. A spatiotemporal input data structure linking input target coordinates computed from contemporaneous images to respective reference times is constructed. The spatiotemporal input data structure is processed to identify an input instruction. The identified input instruction is executed on a machine. In another aspect, an image is displayed at a display location disposed between a viewing space and an interactive space, wherein the displayed image is viewable from a perspective in the viewing space. Images of the interactive space are acquired from at least one field of view. An input target is detected in the acquired images.

Abstract: Methods, machines, systems and machine-readable instructions for processing an image are described. In one aspect, the image is divided into a population of image blocks. Frequency domain vectors are generated from respective ones of the image blocks. Each of the frequency domain vectors includes a respective set of values corresponding to a set of transform coefficients. Preliminary estimates of quantization parameter values are determined from frequency distributions of the transform coefficient values in a set of the frequency domain vectors corresponding to a variable sample of the population. The variable sample is determined at least in part by at least one threshold. Values of quantization parameters are estimated from the preliminary estimates.

Abstract: Various method and system embodiments of the present invention are directed to computational estimation of a tempo for a digitally encoded musical selection. In certain embodiments of the present invention, described below, a short portion of a musical selection is analyzed to determine the tempo of the musical selection. The digitally encoded musical selection sample is computationally transformed to produce a power spectrum corresponding to the sample, in turn transformed to produce a two-dimensional strength-of-onset matrix. The two-dimensional strength-of-onset matrix is then transformed into a set of strength-of-onset/time functions for each of a corresponding set of frequency bands. The strength-of-onset/time functions are then analyzed to find a most reliable onset interval that is transformed into an estimated tempo returned by the analysis.

Abstract: Systems and methods of reducing artifacts in compressed images are described. In one aspect, spatially-shifted forward transforms of the input image are computed to generate respective sets of forward transform coefficients. Nonlinear transforms are applied to the forward transform coefficients of each set. Inverse transforms of the sets of nonlinearly transformed forward transform coefficients are computed to generate respective intermediate images. Respective measures of local spatial intensity variability are computed for pixels of each of the intermediate images. An output image is computed with pixel values computed based at least in part on the computed measures of local spatial intensity variability.

Abstract: An exemplary method for providing an optimized digital image comprises obtaining digital image data by a digital image capture device, based on the digital image data, automatically determining at least two setting types, and a plurality of setting values therefor, for capturing a plurality of digital images, automatically capturing the plurality of digital images using the digital image capture device, each image being captured with at least one setting value different than another image, and providing an optimized digital image based on the plurality of digital images. In one exemplary embodiment, the optimized image is selected from the plurality of digital images. In another exemplary embodiment, the optimized image is synthesized based on the plurality of digital images.

Abstract: A technique of reconstructing a high resolution image from at least one image sequence of temporally related high and low resolution image frames wherein each of said high resolution image frames includes a low spatial frequency component and a high spatial frequency component is described. The high-resolution image reconstruction technique uses spatial interpolation to generate a low spatial frequency component from a low-resolution image frame of the image sequence. The technique is adapted to generate a high spatial frequency component from at least one high resolution image frame of the image sequence which is closely related to the low resolution image frame, and to remap the high spatial frequency component to a motion-compensated high spatial frequency component estimate of the low resolution image frame.

Abstract: Methods, machines, systems and machine-readable instructions for processing input audio signals are described. In one aspect, an input audio signal has a noise period that includes a targeted noise signal and a noise-free period free of the targeted noise signal. The input audio signal in the noise-free period is divided into spectral time slices each having a respective spectrum. Ones of the spectral time slices of the input audio signal are selected based on the respective spectra of the spectral time slices. An output audio signal is composed for the noise period based at least in part on the selected ones of the spectral time slices of the input audio signal in the noise-free period.

Abstract: A “path-enhanced” recorder combines the recording of multimedia data files (typically, but not necessarily, using conventional digital audio, photo, and video recording technology), along with quasi-continuous capturing of time and position data to permit a subsequent reconstruction of the user's entire path, preferably including times and/or locations at which no multimedia data files were being recorded. A “path-enhanced” recorder device thus provides a combined record of not only multimedia data files but also data concerning path traveled between (and possibly during) the capturing of individual recorded media files to thereby create “path-enhanced” multimedia, which in turn facilitates a number of contemplated applications including the Automatic Generation of Presentations; Exploration, Enhancement and Editing of “Path-Enhanced” Multimedia; and Searching of “Path-Enhanced” Multimedia Databases.

Abstract: In one aspect, audio summaries and transition audio segments are sequentially rendered with at least one transition audio segment rendered between each pair of sequential audio summaries. Each audio summary comprises digital content summarizing at least a portion of a respective associated audio piece. In another aspect, an original audio file is annotated by embedding therein information enabling rendering of at least one audio summary contained in the annotated audio file and comprising digital content summarizing at least a portion of the original audio file. In another aspect, an original audio file is annotated by providing at least one browsable link between the original audio file and at least one audio summary comprising digital content summarizing at least a portion of the original audio file, and storing the original audio file, the at least one browsable link, and the at least one audio summary on a common portable storage medium. In another aspect, an audio piece is divided into audio segments.

Abstract: The present invention provides a method of and system for browsing of music. In an embodiment, a method of browsing recorded music comprises steps of: selecting a song from a library; playing at least a portion of the selected song for a user; while the portion of the selected song is playing, accepting input from the user, the input comprising an indication of the user's enjoyment of the at least a portion of the selected song; repeating said steps of selecting, playing and accepting to generate a sequence of song portions; and creating a record comprising an identification of each selected song portions and the indication for the song portions.