Patent application number | Description | Published |
20110029312 | METHODS AND SYSTEMS FOR ADAPTING A MODEL FOR A SPEECH RECOGNITION SYSTEM - Methods are disclosed for identifying possible errors made by a speech recognition system without using a transcript of words input to the system. A method for model adaptation for a speech recognition system includes determining an error rate, corresponding to either recognition of instances of a word or recognition of instances of various words, without using a transcript of words input to the system. The method may further include adjusting an adaptation, of the model for the word or various models for the various words, based on the error rate. Apparatus are disclosed for identifying possible errors made by a speech recognition system without using a transcript of words input to the system. An apparatus for model adaptation for a speech recognition system includes a processor adapted to estimate an error rate, corresponding to either recognition of instances of a word or recognition of instances of various words, without using a transcript of words input to the system. The apparatus may further include a controller adapted to adjust an adaptation of the model for the word or various models for the various words, based on the error rate. | 02-03-2011 |
20110029313 | METHODS AND SYSTEMS FOR ADAPTING A MODEL FOR A SPEECH RECOGNITION SYSTEM - Methods are disclosed for identifying possible errors made by a speech recognition system without using a transcript of words input to the system. A method for model adaptation for a speech recognition system includes determining an error rate, corresponding to either recognition of instances of a word or recognition of instances of various words, without using a transcript of words input to the system. The method may further include adjusting an adaptation, of the model for the word or various models for the various words, based on the error rate. Apparatus are disclosed for identifying possible errors made by a speech recognition system without using a transcript of words input to the system. An apparatus for model adaptation for a speech recognition system includes a processor adapted to estimate an error rate, corresponding to either recognition of instances of a word or recognition of instances of various words, without using a transcript of words input to the system. The apparatus may further include a controller adapted to adjust an adaptation of the model for the word or various models for the various words, based on the error rate. | 02-03-2011 |
20140278391 | APPARATUS AND METHOD TO CLASSIFY SOUND TO DETECT SPEECH - Audio frames are classified as either speech, non-transient background noise, or transient noise events. Probabilities of speech or transient noise event, or other metrics may be calculated to indicate confidence in classification. Frames classified as speech or noise events are not used in updating models (e.g., spectral subtraction noise estimates, silence model, background energy estimates, signal-to-noise ratio) of non-transient background noise. Frame classification affects acceptance/rejection of recognition hypothesis. Classifications and other audio related information may be determined by circuitry in a headset, and sent (e.g., wirelessly) to a separate processor-based recognition device. | 09-18-2014 |
20150039309 | METHODS AND SYSTEMS FOR IDENTIFYING ERRORS IN A SPEECH REGONITION SYSTEM - Methods are disclosed for identifying possible errors made by a speech recognition system without using a transcript of words input to the system. A method for model adaptation for a speech recognition system includes determining an error rate, corresponding to either recognition of instances of a word or recognition of instances of various words, without using a transcript of words input to the system. The method may further include adjusting an adaptation, of the model for the word or various models for the various words, based on the error rate. Apparatus are disclosed for identifying possible errors made by a speech recognition system without using a transcript of words input to the system. An apparatus for model adaptation for a speech recognition system includes a processor adapted to estimate an error rate, corresponding to either recognition of instances of a word or recognition of instances of various words, without using a transcript of words input to the system. The apparatus may further include a controller adapted to adjust an adaptation of the model for the word or various models for the various words, based on the error rate. | 02-05-2015 |
20150302853 | APPARATUS AND METHOD TO CLASSIFY SOUND TO DETECT SPEECH - Audio frames are classified as either speech, non-transient background noise, or transient noise events. Probabilities of speech or transient noise event, or other metrics may be calculated to indicate confidence in classification. Frames classified as speech or noise events are not used in updating models (e.g., spectral subtraction noise estimates, silence model, background energy estimates, signal-to-noise ratio) of non-transient background noise. Frame classification affects acceptance/rejection of recognition hypothesis. Classifications and other audio related information may be determined by circuitry in a headset, and sent (e.g., wirelessly) to a separate processor-based recognition device. | 10-22-2015 |
Patent application number | Description | Published |
20130240413 | ADJUSTABLE MAGNETIC SEPARATOR - An adjustable magnetic separator and method is provided herein to process aggregate material including magnetic material into a magnetic portion having a predetermined magnetic susceptibility, and a non-magnetic portion. The method and system include configuring an adjustable magnet to provide an effective magnetic field at a separating surface of the magnetic separator corresponding to the predetermined magnetic susceptibility of the magnetic portion to be separated. The strength or intensity of the effective magnetic field may be varied by mechanically adjusting the position of the magnet array included in the adjustable magnet relative to a separating surface defined by the magnetic separator. | 09-19-2013 |
20130240415 | METHOD AND SYSTEM FOR PROCESSING SLAG MATERIAL - A method and system is used to process slag material to yield by-products including a finished iron rich product and a finished low iron fines product. The by-products may include a finished high iron product and a finished medium iron product. The method and system include size classifying the material into a plurality of sized groups prior to using magnetic separation to separate at least one of the sized groups into two portions having differing magnetic susceptibilities. The method and system may include more than one phase of size classifying the material into a plurality of sized groups and using magnetic separation to separate at least one of the sized groups into portions, where the average size of the material remaining after one phase is reduced prior to the subsequent phase. | 09-19-2013 |
20130263890 | CLEANING OF RADIOACTIVE CONTAMINATION USING DRY ICE - A system and method for cleaning components contaminated with radioactive materials uses dry ice blasting to remove the radioactive contaminant from the contaminated component in a chamber including a directed air flow. The removed contaminant is transported by the directed air flow for collection in a filter, which may be configured as a HEPA filter. During the cleaning process, the dry ice used for the blasting process sublimates into gaseous carbon dioxide such that no incremental contaminated cleaning media or residue is generated during decontamination of the component. Because the dry ice is non-corrosive, non-abrasive and environmentally neutral, damage to the component being cleaned and decontaminated is prevented or substantially avoided. Because dry ice sublimates on impact, cleaning of electrical components without compromising electrical function is enabled. | 10-10-2013 |
20150209798 | METHOD AND SYSTEM FOR PROCESSING SLAG MATERIAL - A method and system is used to process slag material to yield by-products including a finished iron rich product and a finished low iron fines product. The by-products may include a finished high iron product and a finished medium iron product. The method and system include size classifying the material into a plurality of sized groups prior to using magnetic separation to separate at least one of the sized groups into two portions having differing magnetic susceptibilities. The method and system may include more than one phase of size classifying the material into a plurality of sized groups and using magnetic separation to separate at least one of the sized groups into portions, where the average size of the material remaining after one phase is reduced prior to the subsequent phase. | 07-30-2015 |