Abstract: This disclosure relates to a stretchable composite electrode and a fabricating method thereof, and particularly relates to a stretchable composite electrode including a silver nanowire layer and a flexible polymer film and a fabricating method thereof.

Abstract: A health care device and a health care method are illustrated, the health care method is used to arrange the health care device at a predetermined location in front of a belly button of a user with a predetermined distance, and to emit the low-frequency wave with a predetermined frequency to the belly button by using a low-frequency wave emitter. The health care device is formed by the low-frequency wave emitter and a cone/pyramid part. The predetermined frequency is 1.27 Hz to 1.81 Hz, and the predetermined distance is 5 cm to 8 cm. The above health care device and method can increase contents of active T cells and B cells in blood and increase an ability of NK cell strains for poisoning cancer cell strains (K562). In short, the above health care device and method have health benefits without a risk of excessive energy causing harm.

Abstract: Systems, methods, and computer-readable storage devices are disclosed for improved real-time audio processing. One method including: receiving audio data including a plurality of frames having a plurality of frequency bins; calculating, for each frequency bin, an approximate speech signal estimation based on the plurality of frames; calculating, for each approximate speech signal estimation, a clean speech estimation and at least one additional target including an ideal ratio mask using a trained neural network model; and calculating, for each frequency bin, a final clean speech estimation using the calculated at least one additional target including the calculated ideal ratio mask and the calculated clean speech estimation.

Abstract: Systems, methods, and computer-readable storage devices are disclosed for improved real-time audio processing. One method including: receiving audio data including a plurality of frames having a plurality of frequency bins; calculating, for each frequency bin, an approximate speech signal estimation based on the plurality of frames; calculating, for each approximate speech signal estimation, a clean speech estimation and at least one additional target including an ideal ratio mask using a trained neural network model; and calculating, for each frequency bin, a final clean speech estimation using the calculated at least one additional target including the calculated ideal ratio mask and the calculated clean speech estimation.

Abstract: A method for sign language recognition includes receiving, with an input device, an input based on a plurality of hand movements and postures of a user that correspond to a sequence of signs, extracting a plurality of features from the input corresponding to the plurality of hand movements and postures, identifying, a start of the sequence of signs in the input based on a first set of features in the plurality of features and a first Hidden Markov Model (HMM) stored in the memory and identifying a first sign in the input based on a second set of features in the plurality of features and a second HMM stored in the memory. The method also includes generating an output corresponding to the first sign from the input.

Abstract: A method for sign language recognition includes receiving, with an input device, an input based on a plurality of hand movements and postures of a user that correspond to a sequence of signs, extracting a plurality of features from the input corresponding to the plurality of hand movements and postures, identifying, a start of the sequence of signs in the input based on a first set of features in the plurality of features and a first Hidden Markov Model (HMM) stored in the memory and identifying a first sign in the input based on a second set of features in the plurality of features and a second HMM stored in the memory. The method also includes generating an output corresponding to the first sign from the input.

Abstract: A method of interpreting input from a user includes providing a surface within reach of a hand of the user. A plurality of locations on the surface that are touched by the user are sensed. An alphanumeric character having a shape most similar to the plurality of touched locations on the surface is determined. The determining includes collecting information associated with hand region localized modules, and modeling the information using statistical models. The user is informed of the alphanumeric character and/or a word in which the alphanumeric character is included. Feedback is received from the user regarding whether the alphanumeric character and/or word is an alphanumeric character and/or word that the user intended to be determined in the determining step.

Abstract: An in-vehicle information system includes a camera and a controller that accept gesture input. A controller receives frames of video data and generates trajectory data for a movement of a hand in the video data. The controller uses a first hidden Markov model (HMM) to decode a sequence of strokes from the trajectory data, removes a starting and ending stroke to form an edited stroke sequence, and re-normalizes the strokes in the edited stroke sequence. The controller uses a second HMM corresponding to a predetermined set of characters to identify a character corresponding to the re-normalized edited stroke sequence.

Abstract: An in-vehicle information system includes a camera and a controller that accept gesture input. A controller receives frames of video data and generates trajectory data for a movement of a hand in the video data. The controller uses a first hidden Markov model (HMM) to decode a sequence of strokes from the trajectory data, removes a starting and ending stroke to form an edited stroke sequence, and re-normalizes the strokes in the edited stroke sequence. The controller uses a second HMM corresponding to a predetermined set of characters to identify a character corresponding to the re-normalized edited stroke sequence.

Abstract: A method and apparatus for performing discriminative training of, for example, call routing training data (or, alternatively, other classification training data) which improves the subsequent classification of a user's natural language based requests. An initial scoring matrix is generated based on the training data and then the scoring matrix is adjusted so as to improve the discrimination between competing classes (e.g., destinations). In accordance with one illustrative embodiment of the present invention a Generalized Probabilistic Descent (GPD) algorithm may be advantageously employed to provide the improved discrimination.

Abstract: A method and system of identifying and spotting segments containing key information in voice messages. The method can be used to spot a key segment such as a name segment in a voice message by detecting and verifying the presence of a phrase such as “My name is . . . ” or “This is . . . ”. Once the key segment of interest has been spotted, the method provides the user with only the pertinent information (e.g., the name of the caller), which is contained in the key segment. This allows a user retrieving a message to hear just a desired section or sections of a message without listening to the rest of the message.

Abstract: A method and apparatus for performing real-time endpoint detection for use in automatic speech recognition. A filter is applied to the input speech signal and the filter output is then evaluated with use of a state transition diagram (i.e., a finite state machine). The filter is advantageously designed in light of several criteria in order to increase the accuracy and robustness of detection. The state transition diagram advantageously has three states. The endpoints which are detected may then be advantageously applied to the problem of energy normalization of the speech portion of the signal.

Abstract: A system for, and method of, measuring a degree of independence of semantic classes in separate domains. In one embodiment, the system includes: (1) a cross-domain distance calculator that estimates a similarity between n-gram contexts for the semantic classes in each of the separate domains to determine domain-dependent relative entropies associated with the semantic classes and (2) a distance summer, associated with the cross-domain distance calculator, that adds the domain-dependent distances over a domain vocabulary to yield the degree of independence of the semantic classes.

Abstract: A speaker verification method and apparatus which advantageously minimizes the constraints on the customer and simplifies the system architecture by using a speaker dependent, rather than a speaker independent, background model, thereby obtaining many of the advantages of using a background model in a speaker verification process without many of the disadvantages thereof. In particular, no training data (e.g. speech) from anyone other than the customer is required, no speaker independent models need to be produced, no a priori knowledge of acoustic rules are required, and, no multi-lingual phone models, dictionaries, or letter-to-sound rules are needed. Nonetheless, in accordance with an illustrative embodiment of the present invention, the customer is free to select any password phrase in any language.

Abstract: A method and apparatus for performing real-time endpoint detection for use in automatic speech recognition. A filter is applied to the input speech signal and the filter output is then evaluated with use of a state transition diagram (i.e., a finite state machine). The filter is advantageously designed in light of several criteria in order to increase the accuracy and robustness of detection. The state transition diagram advantageously has three states. The endpoints which are detected may then be advantageously applied to the problem of energy normalization of the speech portion of the signal.

Abstract: A free-standing flow-through capacitor (FTC) is constructed by concentrically winding two electrodes and two dividers into a hollow-center roll. A liquid-feeding pipe is inserted to the central opening for delivering fluids to the FTC. Nanoparticles of hydrated iron compound with Fe3O4 as the main component or its composite powders are used as the active materials for the electrodes. With channels crated by the dividers assembled in the roll, fluids injected from the feed pipe are confined inside the FTC, and flow outwardly and transversely through the entire length of the electrodes. Under an application of a low DC voltage to the electrodes, charged species are adsorbed and removed from the treated liquids as soon as they are in contact with the electrodes. Capacitive deionization using FTC of the present invention is applicable to waste-streams reduction, water purification and desalination at low costs and easy operation.

Abstract: A method and apparatus for performing discriminative training of, for example, call routing training data (or, alternatively, other classification training data) which improves the subsequent classification of a user's natural language based requests. An initial scoring matrix is generated based on the training data and then the scoring matrix is adjusted so as to improve the discrimination between competing classes (e.g., destinations). In accordance with one illustrative embodiment of the present invention a Generalized Probabilistic Descent (GPD) algorithm may be advantageously employed to provide the improved discrimination.

Abstract: A customized method or algorithm for holding an interactive dialogue session between a (human) user and a machine (hereinafter referred to simply as a “dialogue”) is generated, such that the resulting dialogue advantageously responds to the user's requests and wherein the system's capability (i.e., the dialogue) is automatically modified thereafter based on dynamically changing external databases. Specifically, a computer system acts as a Dialogue Generator agent by creating such a customized dialogue consisting of services that are organized and presented in a form that is a combination of the user's expectations and the system's capabilities.

Abstract: The present invention provides a method of forming a node contact hole on a semiconductor wafer. The semiconductor wafer comprises a silicon substrate, a first dielectric layer positioned on the silicon substrate, two bit lines positioned on the first dielectric layer which form a first groove between the two bit lines and the surface of the first dielectric layer, and a second dielectric layer positioned on each of the two bit lines. A lithographic process is performed to form a photoresist layer on the second dielectric layer with at least one second groove extending down to the second dielectric layer wherein the second groove is positioned above the first groove and is perpendicular to the first groove. An etching process is performed along the second groove of the photoresist layer to remove the second dielectric layer and the first dielectric layer under the second groove down to the surface of the silicon substrate so as to approximately form the node contact hole.

Abstract: The present invention provides a method of forming a passivation layer of a DRAM on a semiconductor wafer. The semiconductor wafer comprises a silicon substrate, and two adjacent gates positioned on the surface of the silicon substrate wherein each gate comprises a spacer on each of two opposite walls. The method comprises forming a first silicon-oxygen-nitride layer of predetermined thickness on the semiconductor wafer, forming a second silicon-oxygen-nitride layer of predetermined thickness on the first silicon-oxygen-nitride layer and forming a BPSG (borophosphosilicate glass) layer uniformly on the second silicon-oxygen-nitride to planarize the surface of the semiconductor wafer. The BPSG layer is used as a dielectric layer. The first silicon-oxygen-nitride layer serves as diffusion barrier layer to prevent diffusion of boron and phosphorous from the BPSG layer into the silicon substrate. The second silicon-oxygen-nitride layer is used as an etching stop layer.