Loading patent…
AssigneeLoquendo S.p.A.
InventorsRoberto Gemello, Franco Mana
Estimating the term…
A method for extracting sampled voice signal features for an automatic voice recognition system essentially comprises the following steps: decomposing the sampled voice signal to obtain decomposition of the signal into a plurality of subbands by means of a digital bank of filters whose structure is that of a fully developed, symmetric binary tree (20), performing a discrete wavelet transform, each node (21, 23, . . . ) of the binary tree being associated to one of the subbands; employing all the subbands obtained by means of the binary tree (20) to generate the corresponding parameters representing the features extracted from the sampled voice signal.
Method for extracting sampled voice signal features (S) for an automatic voice recognition system (S), characterised in that it comprises the following steps:
Method as per any of the preceding claims, in which said binary tree structure comprises six levels.
Dependencies are inferred from the claim text.
Drawings are temporarily unavailable. Open the PDF at USPTO
No related US applications are recorded in the indexed filing history.
Backward citations are indexed for grants only.
Speech recognition: feature extraction for speech recognition; Selection of recognition unit
Speech analysis techniques or speech synthesis
Term end not estimated
The invention refers to automatic voice recognition systems in general and specifically refers to a method for extracting voice signal features for recognition.
An automatic voice recognition procedure may be schematically illustrated as a plurality of modules arranged in series between a voice signal input and a recognised sequence of words output:
a first signal processing module acquires the input voice signal transforming it from analogue to digital and sampling it as required;
Additional characteristics and advantages of the invention will now be described, by way of example only, with reference to the accompanying drawings wherein:
FIG. 1 is a block diagram of a method for extracting voice signal features according to the invention;
FIG. 2 is a schematic diagram that illustrates a filtering and subsampling tree; and
A method for extracting the features of a sampled voice signal S will now be described in detail with reference to FIG. 1. FIG. 1 shows a block diagram of an acoustic front-end, of the wavelet or MRA (Multi Resolution Analysis) type, made according to the invention. The diagram in FIG. 1 comprises a sequence of seven blocks, from a first block 2, where a sampled voice signal S is input, to a block 14, which outputs the extracted features C.
The sampled voice signal S is obtained by means of acquiring and sampling means, implemented for example as an acquiring and sampling unit not shown in FIG. 1.
We will now analyse the seven blocks forming the wavelet front-end shown in FIG. 1 in detail.