Kabushiki Kaisha Toshiba

Tokyo Shibaura Denki Kabushiki Kaisha

Toshiba Corporation

Tsuneo Nitta

Sadakazu Watanabe

Kensuke Uehara

Teruhiko Ukita

Yasuo Takahashi

Kenichi Maeda

Hiroshi Matsuura

Hidenori Shinoda

Tomio Sakata

Yoshiyuki Hara

Hideki Kasuya

Takashi Kondo

Yoichi Takebayashi

Hideki Satoh

Masatoshi Tanaka

Hiroyuki Tsuboi

Toshitake Murata

Haruo Asada

Shouichi Hirai

A sound analyzer sound analyzes an input speech signal to obtain feature vectors. A matrix quantizer performs a matrix quantization process between the feature vectors obtained by the sound analyzer and a phonetic segment dictionary prepared in phonetic segment units to obtain a phonetic segment similarity sequence. A PS-phoneme integrating section integrates the phonetic segment similarity sequence into a phonemic feature vector. A HMM recognizer checks the phonemic feature vector using a HMM prepared in certain units, to thereby perform a recognition process.

Speech recognition system and method which permits a speaker's utterance

Synthesized speech is generated by a software-implemented system with a programmed central processing unit. Phonetic parameters are generated from a series of phonetic symbols of an input text to be converted into synthesized speech, and prosodic parameters are also generated from prosodic information of the input text. The activity ratio of the central processing unit is determined, and the order of phonetic parameters or the arrangement of a synthesis unit or filter for speech synthesis is determined depending on the determined activity ratio of the central processing unit. Synthesized speech sounds are generated and filtered based on the phonetic and prosodic parameters according to the determined order of phonetic parameters or the determined arrangement of the filter.

Text-to-speech synthesis with controllable processing time and speech

A recognition system comprises a feature extractor for extracting a feature vector x from an input speech signal, and a recognizing section for defining continuous density Hidden Markov Models of predetermined categories k as transition network models each having parameters of transition probabilities p(k,i,j) that a state Si transits to a next state Sj and output probabilities g(k,s) that a feature vector x is output in transition from the state Si to one of the states Si and Sj, and recognizing the input signal on the basis of similarity between a sequence X of feature vectors extracted by the feature extractor and the continuous density HMMs. Particularly, the recognizing section includes a memory section for storing a set of orthogonal vectors .phi..sub.m (k,s) provided for the continuous density HMMs, and a modified CDHMM processor for obtaining each of the output probabilities g(k,s) for the continuous density HMMs in accordance with corresponding orthogonal vectors .phi..sub.m (k,s).

Speech recognition using continuous density hidden markov models and the

A speech detection apparatus capable of reliably detecting speech segments in audio signals regardless of the levels of input audio signals and background noises. In the apparatus, a parameter of input audio signals is calculated frame by frame, and then compared with a threshold in order to judge each input frame as one of a speech segment and a noise segment, while the parameters of the input frames judged as the noise segments are stored in the buffer and the threshold is updated according to the parameters stored in the buffer. The apparatus may utilize a transformed parameter obtained from the parameter, in which the difference between speech and noise is emphasized, and noise standard patterns are constructed from the parameters of the input frames pre-estimated as noise segments.

Speech detection apparatus not affected by input energy or background

An inner product computing unit computes inner products of an input pattern whose category is unknown, and orthogonalized dictionary sets of a plurality of reference patterns whose categories are known. A nonlinear converting unit nonlinearly converts the inner products in accordance with a positive-negative symmetrical nonlinear function. A neural network unit or a statistical discriminant function computing unit performs predetermined computations of the nonlinearly converted values on the basis of preset coefficients in units of categories using a neural network or a statistical discriminant function. A determining section compares values calculated in units of categories using the preset coefficients with each other to discriminate a category to which the input pattern belongs.

Pattern recognition system and method using neural network

A plurality of candidate phonetic segments extracted from the input speech signal are passed through transition networks prepared for the respective words so as to obtain a score by weighting/averaging the long-term strategic scores by taking consideration of statistic distribution of the similarities or distances of phonetic segments and the short-term strategic scores by taking consideration of the environment of the phonetic segments.

Speech recognition system utilizing both a long-term strategic and a

An orthogonalizing time filter section is arranged in place of a Gram Schmidt orthogonalizing section. The orthogonalizing time filter section is constituted by a plurality of filters for performing smoothing processing and differential processing. The orthogonalizing time filter section obtains an average pattern of acquired learning patterns, and smoothes the average pattern along the time base to obtain a dictionary of a first axis. The section differentiates the average pattern along the time base to obtain a dictionary of a second axis. The above processing is repeated for each category, thus generating an orthogonalized dictionary.

Speech recognition apparatus and method utilizing an orthogonalized

Speech pattern data representing speech of a plurality of speakers are stored in a pattern storage section in advance. Averaged pattern data obtained by averaging a plurality of speech pattern data of the first of the plurality of speakers are obtained. Data obtained by blurring and differentiating the averaged pattern data are stored in an orthogonalized dictionary as basic orthogonalized dictionary data of first and second axes, respectively. Blurred data and differentiated data obtained with respect to the second and subsequent of the plurality of speakers are selectively stored in the orthogonalized dictionary as additional dictionary data having new axes. Speech of the plurality of speakers is recognized by computing a similarity between the orthogonalized dictionary formed in this manner and input speech.

Orthogonalized dictionary speech recognition apparatus and method thereof

Phoneme feature parameters are extracted from input digital speech signals by means of LPC analysis. Phonetic segments having phonetical meanings are obtained together with similarities to prescribed basic phonetic segments from the feature parameters to be passed through nodes of transition networks provided for each word. In passing the nodes, scores for similarity Sj of predetermined segments of the corresponding phonetic segments are made in selective scoring and the accumulation of the scores is used for recognition of continuous word speech.

System for continuous speech recognition through transition networks

In a speech recognition system for recognizing speeches uttered by non-specific speakers, start and end points of a word or speech interval are determined by a novel preprocessor for searching a sound power level to obtain speech boundary candidates and for determining likelihoods of speech or word intervals on the basis of the boundary candidates. Since likelihoods (probabilities) are determined for speech interval candidates, the similarity rate between feature parameters and reference pattern set of a speech signal are calculated for only the higher likelihood candidates, thus improving the accuracy and the speed of speech recognition. A percentage of erroneous boundary decision is about 0.5% when two speech interval candidates of the first and second likelihoods are adopted.

Speech recognition system

An IC card used for identifying an individual, has a key pad on the surface, a memory for storing a program and data, and a planar microphone and speaker. Speech corresponding to data or a command is converted into signals by a microphone. These signals can be supplied to an output terminal of the IC card under control of the IC card. Alternatively, the IC card can store or analyze the speech and generate a corresponding command or data. As the number of functions of the IC card is increased, the number of commands must also be increased. By inputting commands or data by voice instead of dedicated function keys, the number of necessary keys is reduced and convenience is enhanced.

IC card

A continuous speech recognition circuit has a data generating circuit for calculating feature pattern data each having N-frame feature parameter data of a plurality of word-periods and reference pattern data every time one-frame period has elapsed and for sequentially generating a maximal similarity data among the calculated similarity data, and a recognition circuit for detecting a series of continuous word-periods which gives the largest similarity sum within a speech interval in accordance with the similarity data from the data generating circuit and recognizing as effective word data the word series corresponding to the detected series of continuous word-periods. The similarity data in each word period is obtained by calculating partial similarity data between the feature parameter data of each frame and each reference parameter data and using the N partial similarity data obtained during the word-period.

Continuous speech recognition apparatus

Continuous speech signal is recognized using 'rough' and 'detail' parameters derived from prestored reference speech and current unknown speech. The detail parameters are 16 spectral coefficients, the rough parameters 2 or 4 spectral coefficients representing the signal. A word interval detector decides segmentation based on rough parameter similarity.

In a pattern recognition system for speech or print, a first memory stores predetermined reference vectors. A second memory stores subsequently-determined reference vectors subsequent to misrecognition when a new speaker or font is inputted, whereby only the deformations (differences) from a predetermined category of vectors are stored.

Pattern recognition apparatus and method for making same

A monosyllabic recognition apparatus is disclosed which includes a first memory which stores reference vowel patterns respectively representing vowel categories of known reference monosyllables of a preselected language, which are classified in accordance with the type of vowel characteristics of the language, the vowel categories independently including categories corresponding to a contracted sound and a syllabic nasal sound in addition to categories of basic vowels; a second memory which stores reference consonant patterns respectively representing consonant categories of the language, which are classified in accordance with the type of consonant characteristics of the language. A characteristic extracting section generates the acoustic parameter data of the input speech, which is divided by a segment processing section into monosyllabic acoustic parameter components. A speech recognition processing section compares each of the monosyllabic acoustic parameter components with the reference vowel and consonant patterns and performs monosyllable recognition in accordance with the determination of the vowel and consonant categories of the reference patterns which maximally resemble the monosyllabic acoustic parameter components.

Growing community of inventors

Yokohama, Japan

Tsuneo Nitta