Abstract: A multi-state pattern recognition model with non-uniform kernel allocation is formed by setting a number of states for a multi-state pattern recognition model and assigning different numbers of kernels to different states. The kernels are then trained using training data to form the multi-state pattern recognition model.

Abstract: A reliable full covariance matrix estimation algorithm for pattern unit's state output distribution in pattern recognition system is discussed. An intermediate hierarchical tree structure is built to relate models for product units. Full covariance matrices of pattern unit's state output distribution are estimated based on all the related nodes in the tree.

Abstract: A multimedia play television is proposed, wherein a central processor, a television signal receiver, a memory card slot, and an optical disc read/write device connected with the central processor are integrated together. The central processor processes television signals received by the television signal receiver and plays them on the television. The central processor controls actions of the memory card slot and the optical disc read/write device to let read data in a memory card or an optical disc be played on the television. The proposed television can read memory cards and optical discs, and makes use of a single central processor to process multimedia audio/video data to accomplish multiple selections of entertainments, convenient use and less occupied space.

Abstract: A method and apparatus for extracting speech features from a speech signal in which the linear frequency spectrum data, as generated, for example, by a conventional frequency transform, is first converted to logarithmic frequency spectrum data having frequency data distributed on a substantially logarithmic (rather than linear) frequency scale. Then, a plurality of digital auditory filters is applied to the resultant logarithmic frequency spectrum data, each of these filters having a substantially similar shape, but centered at different points on the logarithmic frequency scale. Because each of the filters have a similar shape, the feature extraction approach of the present invention advantageously can be easily modified or tuned by adjusting each of the filters in a coordinated manner, with the adjustment of only a handful of filter parameters.

Abstract: A method and apparatus for extracting speech features from a speech signal in which the linear frequency spectrum data, as generated, for example, by a conventional frequency transform, is first converted to logarithmic frequency spectrum data having frequency data distributed on a substantially logarithmic (rather than linear) frequency scale. Then, a plurality of digital auditory filters is applied to the resultant logarithmic frequency spectrum data, each of these filters having a substantially similar shape, but centered at different points on the logarithmic frequency scale. Because each of the filters have a similar shape, the feature extraction approach of the present invention advantageously can be easily modified or tuned by adjusting each of the filters in a coordinated manner, with the adjustment of only a handful of filter parameters.

Abstract: A remote viewing system is for viewing digital images of remote locations. The viewing system includes a plurality of digital image acquisition devices. The devices are located at remote locations. The system also includes a plurality of digital image transmission devices connected to respective ones of the digital image acquisition devices. A digital image receiving device is communicatively connected to each of the digital image transmission devices. A digital image server device is connected to the digital image receiving device. The digital image server device is connected to a network, such as the Internet. A network-enabled computer can access select ones of the digital images over the network. The digital image receiving device and the digital image server device cooperate to make available to the network-enabled computer the select ones of the digital images for download from the network.

Abstract: Speech recognition in which the log probabilities of the null and alternative hypothesis are computed for an input speech sample by comparison with specific stored speech vocabularies/grammars and with general speech characteristics. The difference in probabilities is normalized by the magnitude of the null hypothesis to derive a likelihood factor which is compared with a rejection threshold that is utterance-length dependent. Advantageously, a high-order polynomial representation of the rejection threshold length dependency may be simplified by a series of piece-wise constants which are stored as rejection thresholds to be selected in accordance with the length of the input speech sample.

Abstract: The present invention provides a novel method of analyzing speech signals in order to reduce the computational power required to perform both speech compression and voice recognition operations. Digital speech signals are provided to a speech analyzer which generates a linear predictive coded (LPC) speech analysis signal that is compatible for use in both the voice recognition circuit and the speech compression circuit. The speech analysis signal is then provided to the compression circuit, which further processes the signal into a form used by an encoder and then the encoder encodes the processed signal. The same speech analysis signal is also provided to a voice recognition circuit, which further processes the signal into a form used by a recognizer and then the recognizer performs recognition on the processed signal.

Abstract: A facility is provided for allowing a caller to place a telephone call by merely uttering a label identifying a desired called destination and to charge the telephone call to a particular billing account by merely uttering a label identifying that account. Alternatively, the caller may place the call by dialing or uttering the telephone number of the called destination or by entering a speed dial code associated with that telephone number. The facility includes a speaker verification system which employs cohort normalized scoring. Cohort normalized scoring provides a dynamic threshold for the verification process making the process more robust to variation in training and verification utterences. Such variation may be caused by, e.g., changes in communication channel characteristics or speaker loudness level.