Sharp Kabushiki Kaisha Corporation

Yoshiji Fujimoto

Toshio Akabane

Naoyuki Fukuda

Atsunori Kitoh

Kenji Sakamoto

Tomokazu Morio

Masayuki Katagiri

Shuichi Kawama

Kouichi Yamaguchi

A voice recognition apparatus capable of recognizing any word utterance by using a neural network, the apparatus includes a unit for inputting an input utterance and for outputting compressed feature variables of the input utterance a unit using a neural network and connected to the input unit for receiving the compressed feature variables output from the input unit and for outputting a value corresponding to a similarity between the input utterance and words to be recognized. The neural network unit has a first unit for outputting a value which corresponds to a similarity in partial phoneme series of a specific word among vocabularies to be recognized with respect to the input utterance. The neural network also has a second unit connected to the first unit for receiving all of the values output from the first unit and for outputting a value corresponding to a similarity in the specific word with respect to the input utterance. And the neural network also has a third unit connected to the second unit for receiving all of the values output from the second unit and for outputting a value corresponding to a classification of voice recognition in which the input utterance belongs.

Speech recognition system with neural network

An apparatus for calculating a sum of products of proper elements in matrix, connected to a receiving device (21) for receiving an input vector and having a memory (11, 12) capable of storing data used to determine whether each of the matrix elements is zero or nonzero and storing data representative of contents of the matrix elements having nonzeros, the apparatus being capable of reducing a required capacity of the memory (11, 12) and a number of calculation processes in a linear transformation of the input vector received at the receiving device (21) by using a transformation matrix having matrix elements of a large scale sparse matrix, the transformation matrix being represented in a two-dimensional array and stored in the memory (11, 12), the apparatus includes a determining unit connected to the memory (11, 12) for determining whether each of the matrix elements in the transformation matrix is zero or nonzero in accordance with the zero data or the nonzero data stored in the memory (11, 12), and a calculating unit (1) connected to the receiving device (21), the determining unit (2), the memory (11, 12) and the outputting device (22) respectively for calculating the sum of products of the nonzero data store in the memory (11, 12) with the input vector with respect to each of the matrix elements which is identified as nonzeros by the determining unit (2).

Apparatus for calculating matrices

A voice recording and reproducing apparatus which includes a voice encoder for encoding a voice signal into a voice code, an address counter, a voice code memory in which the voice code is circulatively stored as specified by the address counter, a discriminator for determining whether the voice signal inputted is a voiced signal or whether the voice signal inputted is a silent signal, and a register for storing an address value of the address counter when the discriminator indicates that the silent signal has changed into the voiced signal. During reproduction, the starting address of the reproduction is determined by a stored address. The stored address represents the address value in the address counter at the time when the discriminator indicates that a silent signal has changed to a voiced signal.

Voice recording and reproducing apparatus

A fingerprint input apparatus is constructed to have simple construction. It serves to provide an excellen S/N ratio and reliably input a fingerprint pattern. The fingerprint input apparatus includes a light source for emitting a ray of light, a lightguide plate having a predetermined location at which a finger can be contact with the lightguide plate, a unit for directing the ray of light into at least one surfaces of the lightguide plate in a manner to keep such an angle of incidence as allowing the ray of light to be totally reflected within the lightguide plate, the lightguide plate being constructed to direct the ray of light into the predetermined location, and an image pickup unit located near the predetermined location of the lightside plate and outside of an opposite surface to a surface having the predetermined location, the image pickup unit being constructed to pick up a right irregularly reflected on a fingerprint surface of the finger contacted with the predetermined location.

Fingerprint input apparatus

A neurocomputer connected to a host computer, the neurocomputer having a plurality of processor elements, each of the processor elements being placed at each of node of a lattice respectively, the neurocomputer includes a plurality of first processor elements, each of the first processor elements being placed at a node of the lattice, capable of transmitting data from and to the host computer and capable of transmitting the data to one of adjacent processor elements, a plurality of second processor elements, each of the second processor elements being placed at a node of the lattice, capable of receiving the data from one of adjacent processor elements, and capable of sending the data to another adjacent processor elements from which the data is not outputted. The neurocomputer also includes a plurality of rectangular regions, each of the rectangular regions including a plurality of the processor elements, a plurality of physical processors, each of the processors being placed in each of the rectangular regions and connected with adjacent processors each other, each of the processors being capable of inputting and outputting to and from the host computer and having all functions of the processor elements included in the rectangular region, and a device for distributing the physical processors to one or a plurality of divided sections formed in the rectangular regions in such a manner that each of the sections is substantially equally assigned to each of the physical processors by permutation.

Neuro-computer

A voice synthesizing device which compiles wave segments, such as pitch wave segments, in order to synthesize speech. Speech is synthesized by connecting wave segments to form a contiguous waveform. Each wave segment is assigned one or more connection types which describe the connection to be made between points on that wave segment and points on adjacent wave segments. A wave segment connector uses information on the connection types of adjacent wave segments to connect the end point and lead point of the adjacent wave segments using a normal sampling period or a normal sampling period compressed or expanded by 1/2 of the sampling period. The period used depends on the connection type stored in the connection type memory.

System and method for reducing distortion in voice synthesis through

A decoder for decoding a delta-modulated code represented with binary digit '1' or '0' to convert it into an analog signal, wherein when the delta-modulated code takes the first value ('1' or '0'), a positive pulse signal is applied to an integrator to increase its accumulated value by a constant value and when the delta-modulated code takes the second value ('0' or '1'), a negative pulse signal is appied to the integrator to decrease its accumulated value by the constant value, and this accumulated value of the integrator is outputted as an analog signal which corresponds to a code train of the delta-modulated code. The decoder is characterized in that it has a pulse width modulater which limits an effective pulse width of the pulse signal by performing pulse width modulation of the positive or negative pulse signal in response to a control signal.

Decoder for delta-modulated code

A speech analyzing and synthesizing apparatus samples a speech signal and its coding section outputs one or more codes of variable length per sampling according to the differential value between two mutually adjacent sampled signals to represent this differential value. The apparatus also includes a decoding section which decodes the outputted codes to obtain the differential value and a limiting circuit for limiting the number of outputtable codes per sample according to the differential value and codes representing differential values previously outputted from the coding section.

Speech analyzing and synthesizing apparatus using reduced number of codes

An image processing system changes the color of a portion of an image to another color while maintaining the shades of the original image and which can also implement color changes between the three primary colors. The system utilizes a calculator which calculates modified intensity information from intensity information of color components read out from an image memory. The modified intensity information or the intensity information read out from the image memory is selected by a selector and fed to a look-up table memory through a gate. The look-up table memory converts the modified intensity information or the intensity information to converted intensity information in accordance with a predetermined function. A display displays an image on the basis of the converted intensity information outputted from the look-up table memory.

Color conversion image processing system with modified intensity

A speech analyzing and synthesizing apparatus includes a sampling circuit for sampling an inputted speech signal, a speech analyzer for analyzing and coding the inputted speech signal, and a speech synthesizer for decoding the speech signal coded by said speech analyzer into a synthesized speech, whereby the average code length can be advantageously shortened to realize a low capacity memory and a well-analyzed and synthetized speech. Moreover, a speech recording apparatus is provided wherein a soundless section in the inputted speech signal is compressed by real time, and even a weak short-time portion in the sounded section of an inputted speech signal is not erroneously judged to be soundless, but is coded as a sounded portion.

Speech analyzing and synthesizing apparatus utilizing differential

A digital filter for obtaining the power of a filtered signal passing through a particular frequency band for each frame (short time division) includes an auto-correlation circuit for calculating the auto-correlation function of the input signal, and a sum of products calculator for obtaining the power of the signal by multiplying a linear coefficient with the auto-correlation function for the frame and summing the products together to get the power of the filtered signal.

Growing community of inventors

Nara, Japan

Yoshiji Fujimoto