US 5,347,612 AGranted

Voice recognition system and method involving registered voice patterns formed from superposition of a plurality of other voice patterns

AssigneeRicoh Company, Ltd.

InventorsJunichiroh Fujimoto, Seigo Yasuda, Tomofumi Nakatani

Life of this patent

Estimating the term…

Loading patent…

Abstract

After sampling a voice signal by a microphone (31), a voice interval is detected by a voice interval detector (32). Then, the voice signal is processed by a frequency analyzer (33) having a predetermined number of channels at a predetermined time interval, whereby a corresponding portion of the voice signal is quantized at each channel. Then, quantized data obtained from these channels are binary converted (34) to thereby form a frame comprised of a series of binary data. Each data of this frame corresponds to one of the channels, and preferably it is set to be an integral multiple of a computer calculation unit (e.g., 4 bits, 8 bits, etc.). When forming a combination frame by superimposing two or more of such frames, the combination frame is divided into layers, whereby it is so structured that each bit is represented by a binary number in each layer. On the other hand, it may also be so structured that one frame is divided in a plurality of sub-frames and a preliminary comparison process is carried out using each sub-frame.

Claims12

Dependent claims

A voice pattern similarity calculating apparatus for calculating a degree of similarity through comparing an input voice pattern and a registered voice pattern, the apparatus comprising:

a) voice pattern generating means for generating a digital voice pattern having a predetermined number of elements by dividing an input voice signal into the same predetermined number of frequency bands and forming a digitized output signal for each of the bands;
b) a library for storing registered voice patterns which include sub-patterns which have been previously formed by a method including the steps of:
1) forming a plurality of voice patterns by repetitively pronouncing the same sound,
2) superimposing the plurality of voice patterns to form a superimposed pattern,
3) digitizing the value of each element of the superimposed pattern to form digitized superimposed pattern values, and
4) forming plural sub-patterns from the digitized superimposed pattern values, the sub-patterns having a plurality of elements, the plural sub-patterns being for forming binary numbers by combining respective elements of the plurality of elements of the sub-patterns;
c) comparison means for comparing:
(1) the predetermined plurality of elements in the digitized voice pattern from the voice pattern generating means, with
(2) the predetermined plurality of elements of the sub-patterns;
and
d) similarity calculating means for calculating the degree of similarity from a result of comparing from the comparison means.

8 more elements

A voice pattern similarity calculating apparatus for calculating a degree of similarity through comparing an input voice pattern and a registered voice pattern, the apparatus comprising:

a) voice pattern generating means for generating a digital voice pattern having a predetermined number of elements by dividing an input voice signal into the same predetermined number of frequency bands and forming a digitized output signal for each of the bands;
b) a library for storing registered voice patterns that include sub-frames which have been previously formed by a method including the steps of:
1) creating a plurality of voice patterns including n frames obtained by pronouncing the same word a number of times;
2) adding the plurality of voice patterns by superimposition, to thereby define a superimposed pattern;
3) assigning a significance-related position for each of the elements in a frame which defines the superimposed pattern;
and
4) collecting elements having the same significance-related position for each frame, to thereby divide the frame into a plurality of sub-frames;
c) comparison means for comparing each element of a digitized voice pattern generated by the voice pattern generating means with each element of a voice pattern registered in the library, wherein for each frame of the digitized voice pattern the frame is made to correspond to a sub-frame in the library;
and
d) similarity calculating means for calculating the degree of similarity from a result of comparing from the comparison means.

7 more elements

Dependencies are inferred from the claim text.