Abstract: A technique for automatically generating thematic summaries for machine readable representations of documents. The technique begins with determining the number of thematic terms to be used based upon the number of thematic sentence to be extracted. To insure some commonality of theme between extracted sentences, the number of thematic terms used should be less than the number of thematic sentences to be extracted. Having determined the appropriate number of thematic terms, next the method identifies the thematic terms within the document. Afterward, each sentence of the document is scored based upon the number of thematic terms contained within the sentence. The desired number of highest scoring sentences are selected as thematic sentences.

Abstract: An unsupervised method of learning the relationships between words and unspecified topics in documents using a computer is described. The computer represents the relationships between words and unspecified topics via word clusters and association strength values, which can be used later during topical characterization of documents. The computer learns the relationships between words and unspecified topics in an iterative fashion from a set of learning documents. The computer preprocesses the training documents by generating an observed feature vector for each document of the set of training documents and by setting association strengths to initial values. The computer then determines how well the current association strength values predict the topical content of all of the learning documents by generating a cost for each document and summing the individual costs together to generate a total cost. If the total cost is excessive, the association strength values are modified and the total cost recalculated.

Abstract: A computer implemented method of preventing the reproduction of specks while reproducing an image including specks represented via source scan lines. According to the method, on a scan like-by-scan line basis, each source scan line signal is transformed into speckless output run lengths. A source run length is selected and examined by a processor to determine whether it represents part of a speck. If the selected source run length might represent a speck, the processor determines whether the selected source run length can be protected from removal. If it cannot, the length of the selected run length and its subsequent neighbor are added to the output run length. As a result, the speck represented by the selected run length disappears from the image by merger with adjacent source run lengths. On the other hand, if the selected source run length is protected from removal, an output run length is output that represents source run length(s) previous to the selected source run length.

Abstract: An iterative method of determining the topical content of a document using a computer. The processing unit of the computer determines the topical content of documents presented to it in machine readable form using information stored in computer memory. That information includes word-clusters, a lexicon, and association strength values. The processing unit beings by generating an observed feature vector for the document being characterized, which indicates which of the words of the lexicon appear in the document. Afterward, the processing unit makes an initial prediction of the topical content of the document in the form of a topic belief vector. The processing unit uses the topic belief vector and the association strength values to predict which words of the lexicon should appear in the document. This prediction is represented via a predicted feature vector.

Abstract: A system and method for unsupervised clustering of audio data segments in an audio data recording containing speech from multiple speakers including the steps of: 1) providing a portion of the audio data containing speech from all of the speakers; 2) forming initial clusters by dividing the portion of the audio data into segments, each of which includes an ordered data set; 3) computing the pairwise distance between each pair of clusters using a likelihood ration independent of the order of data within the segments; and 4) combining into a new cluster the two clusters with a minimum pairwise distance. These steps are repeated until a number of clusters equal to the number of speakers is obtained.

Abstract: A method for segmenting audio data, comprising speech from a plurality of individual speakers, according to speaker is provided. The method comprises providing individual HMMs for each individual speaker, each individual HMM including at least one state, and constructing a speaker network HMM by connecting the individual HMMs in parallel. The audio data is then divided into segments by determining a most likely sequence of states through the speaker network HMM, each of the segments being associated with one of the individual HMMs. Afterward, the speaker of each of the segments is identified. The segmented data may be used to form an index into the audio data according to speaker.

Abstract: The present invention is a digital imaging system for a copier that comprises a full-frame, two-dimensional sensor array to capture a document's image and a full-frame output light valve with backlight to flash expose the image onto a photoreceptor belt. The document is flash exposed to illuminate the page and the full-page image is captured by the full-frame, two dimensional sensor array. The sensor array then reads out the digital image data in response to a driver. The data is then input into the full-frame digital output light valve. The driver provides for the synchronous control of the system. Optionally, an electronic subsystem may perform digital image processing according to user demand--before the data arrives at the output light valve. After image processing (if any), the digital data, supplied to the full-page output light valve, forms a pattern on the display according to the data. The patterned image is then flash exposed by a backlight, through the light valve, and onto a photoreceptor.

Abstract: The present invention describes a method for superimposing prespecified locational, environmental, and contextual controls on user interactions, including interactions of mobile users, with computational resources. A system is described for electronically monitoring contextual information concerning users and machines, including state and locational information including proximity. Interaction policies, including user specified interaction policies, may be registered on an identifiable address path. Methods are described for detecting, selecting and controlling computercontrolled devices, based on the proximity of the device to the user, the current context of the user, the location of other nearby users and devices, and the current state of the devices. Temporary transfer of control, including exclusive control, of particular computers and computer controlled devices to individual users based on the context and environment in proximity to those computing devices is also described.

Abstract: A processor controlled system for correlating an electronic index according to speaker for audio data being recorded in real time. The system includes a source of training data for each of the plurality of individual speakers and audio input system for providing real time audio data including speech for the individual speakers. The audio data is converted into spectral feature data by an audio processor, and is simultaneously recorded on a storage medium by a recording device. A system processor accepts the training data to create individual speaker models, which are combined in parallel to form a speaker network. The system processor then accepts the spectral feature data of the audio data and, using the speaker network, determines segments in the audio data corresponding to each speaker.

Abstract: The present invention describes a method for superimposing prespecified locational, environmental, and contextual controls on user interactions, including interactions of mobile users, with computational resources. A system is described for electronically monitoring contextual information concerning users and machines, including state and locational information including proximity. Interaction policies, including user specified interaction policies, may be registered on an identifiable address path. Methods are described for detecting, selecting and controlling computercontrolled devices, based on the proximity of the device to the user, the current context of the user, the location of other nearby users and devices, and the current state of the devices. Temporary transfer of control, including exclusive control, of particular computers and computer controlled devices to individual users based on the context and environment in proximity to those computing devices is also described.

Abstract: A method for clustering speaker data from a plurality of unknown speakers. The method includes steps of providing a portion of audio data containing speech from at least all the speakers in the audio data and dividing the portion into data clusters. A pairwise distance between each pair of clusters is computed, the pairwise distance being based on a likelihood that two clusters were created by the same speaker, the likelihood measurement being biased by the prior probability of speaker changes. The two clusters with a minimum pairwise distance are combined into a new cluster and speakers models are trained for each of the remaining clusters including the new cluster. The likelihood that two clusters were created by the same speaker may be biased by a Markov duration model based on speaker changes over the length of the initial data clusters.

Abstract: A device for registering component image tiles in a scanning system for transcribing, into electronic form, markings on a two-dimensional surface such as a whiteboard. A registration light pattern is projected onto the surface to be imaged. The pattern projected can be chosen to best suite the specific properties of the camera and imaging environment. The pattern markings are processed using perspective transformations to determine the overlap properties of the image tiles and distortion of each tile image, and the resulting data is used to combine the tiles to produce an undistorted composite image of the whiteboard.

Abstract: In a system of multiple computers connected by a network means, a method for using multiple input styli, each connected to an individual computer, to indicate input information on the display screen of any computer in the system. In a system in which there are many computers, a user of one computer may wish to indicate an input action on a second computer. In the present invention, all pens in a system are able to determine which computer they are writing on and an input position relative to that input computer, regardless of whether they belong to that computer, by determining an identifying "signature" of the computer based on the peak signal strength of a plurality of distinguishable signals in a complex signal produced on the computer display screen and sensed by the stylus. Each stylus is able to communicate with the software entity which can correctly interpret the measurement by using a computer network.

Abstract: The present invention describes a method for superimposing prespecified locational, environmental, and contextual controls on user interactions, including interactions of mobile users, with computational resources. A system is described for electronically monitoring contextual information concerning users and machines, including state and locational information including proximity. Interaction policies, including user specified interaction policies, may be registered on an identifiable address path. Methods are described for detecting, selecting and controlling computercontrolled devices, based on the proximity of the device to the user, the current context of the user, the location of other nearby users and devices, and the current state of the devices. Temporary transfer of control, including exclusive control, of particular computers and computer controlled devices to individual users based on the context and environment in proximity to those computing devices is also described.

Abstract: A full color display is disclosed having a high efficiency light source optically coupled by a light coupling lens array to the active matrix of pixel elements of an image forming means such as a transmissive liquid crystal light valve. The light source is comprised of a patterned matrix of individual, primary colored visible radiation emitters, such as individual phosphor elements, arranged in a manner suitable for the human eye to integrate the individual primary colored lights into a single mixture color. The light coupling lens array collects substantially all of the light rays emitted from each primary colored light emitter and focuses them on a respective one of the transparent electrodes in the active matrix, as a real image smaller than the size of the electrode. A suitable light coupling lens array is a gradient index lens array, a multiple layer, two dimensional array of microlenses functionally analogous to an array of strip lenses, or a microlens array.

Abstract: The present invention describes a method for superimposing prespecified locational, environmental, and contextual controls on user interactions, including interactions of mobile users, with computational resources. A system is described for electronically monitoring contextual information concerning users and machines, including state and locational information including proximity. Interaction policies, including user specified interaction policies, may be registered on an identifiable address path. Methods are described for detecting, selecting and controlling computercontrolled devices, based on the proximity of the device to the user, the current context of the user, the location of other nearby users and devices, and the current state of the devices. Temporary transfer of control, including exclusive control, of particular computers and computer controlled devices to individual users based on the context and environment in proximity to those computing devices is also described.

Abstract: A copyright revenue allocation system that identifies copyrighted works after they are copied. Each copying service within the system informs a copyright revenue allocation service of each page or sheet it copies via document signals representing a portion of the copied document. The copyright revenue allocation service uses the document signals to determine whether the document copied is part of a copyrighted work, and, if so, identifies the owners of the intellectual property contained within that document. After identifying the intellectual property owners, the copyright revenue allocation system increments their associated revenue counters. These revenue counters are used to allocate copyright revenues from the copying services among the intellectual property owners entitled to compensation.

Abstract: A device for transcribing, into electronic form, markings drawn on a whiteboard or blackboard. An electronic camera such as an ordinary video camera is mounted on a computer-controlled pan/tilt head in the ceiling or to the side of the board. Images are captured by directing the camera successively at small regions of the board, then processing these snapshots electronically, leading to a complete, undistorted, high-resolution image of the entire board.

Abstract: The present invention is a novel multilayered structure comprising alternating layers of a base metal and a metal selected from a group of barrier metals. The base metal, in any given layer, is deposited to a thickness less than its critical thickness--a thickness beyond which hillocks are more likely to form for a given temperature. Between each such layer of base metal, a layer of barrier metal is interposed. The intervening layer of barrier metal acts to suppress the formation of hillocks in the base metal.

Abstract: A full color display is disclosed having a high efficiency light source optically coupled by a light coupling lens array to the active matrix of pixel elements of an image forming means such as a transmissive liquid crystal light valve. The light source is comprised of a patterned matrix of individual, primary colored visible radiation emitters, such as individual phosphor elements, arranged in a manner suitable for the human eye to integrate the individual primary colored lights into a single mixture color. The light coupling lens array collects substantially all of the light rays emitted from each primary colored light emitter and focuses them on a respective one of the transparent electrodes in the active matrix, as a real image smaller than the size of the electrode. A suitable light coupling lens array is a gradient index lens array, a multiple layer, two dimensional array of microlenses functionally analogous to an array of strip lenses, or a microlens array.