Technique for modifying reference vector quantized speech feature signals
First Claim
Patent Images
1. An arrangement for modifying a codebook of vector quantized speech feature signals for speaker verification or adaptation to the voice of a particular speaker comprising:
- means for storing a set of vector quantized feature signals q(i), i=1, 2, . . . , K, where K is the number of vector quantized feature signals in said set, and a set of occupancy signals n(i) each corresponding to one of said vector quantized feature signal q(i);
means for receiving an input pattern;
means for analyzing said input pattern to generate a set of input feature vector signals v(t), t=1,2, . . . , M, where M is the number of vector-quantized feature signals in an input pattern;
means responsive to said set of input feature signals and said set of vector quantized feature signals for classifying each input feature signal v(t) as one of said set if vector quantized feature signals q(i);
means responsive to the classification of each input feature signal as one of said vector quantized feature signals for generating an occupancy signal m(i) corresponding to the count of input feature signals classified as each one of said vector quantized feature signals q(i); and
means for modifying said set if vector quantized feature vector signals, q(i), to produce a modified set of vector quantized feature vector signals, q'"'"'(i), based on said classified input feature vector signals, said input feature occupancy signals, said vector quantized feature signals, and said occupancy signals of said vector quantized feature signals
0 Assignments
0 Petitions
Accused Products
Abstract
A signal processing arrangement uses a codebook of first vector quantized speech feature signals formed responsive to a large collection of speech feature signals. The codebook is altered by combining the first speech feature signals of the codebook with second speech feature signals generated responsive to later input speech patterns during normal speech processing. A speaker recognition template can be updated in this fashion to take account of change which may occur in the voice and speaking characteristics of a known speaker.
-
Citations
21 Claims
-
1. An arrangement for modifying a codebook of vector quantized speech feature signals for speaker verification or adaptation to the voice of a particular speaker comprising:
-
means for storing a set of vector quantized feature signals q(i), i=1, 2, . . . , K, where K is the number of vector quantized feature signals in said set, and a set of occupancy signals n(i) each corresponding to one of said vector quantized feature signal q(i); means for receiving an input pattern; means for analyzing said input pattern to generate a set of input feature vector signals v(t), t=1,2, . . . , M, where M is the number of vector-quantized feature signals in an input pattern; means responsive to said set of input feature signals and said set of vector quantized feature signals for classifying each input feature signal v(t) as one of said set if vector quantized feature signals q(i); means responsive to the classification of each input feature signal as one of said vector quantized feature signals for generating an occupancy signal m(i) corresponding to the count of input feature signals classified as each one of said vector quantized feature signals q(i); and means for modifying said set if vector quantized feature vector signals, q(i), to produce a modified set of vector quantized feature vector signals, q'"'"'(i), based on said classified input feature vector signals, said input feature occupancy signals, said vector quantized feature signals, and said occupancy signals of said vector quantized feature signals - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10)
-
5. An arrangement for modifying a codebook of vector quantized feature vector signals according to claim 4 further comprising:
means responsive to said vector quantized feature occupancy signal and said input feature signal occupancy signal for modifying said vector quantized feature occupancy signal.
-
6. An arrangement for modifying a codebook of vector quantized feature vector signals according to claim 5 wherein said vector quantized feature occupancy signal modifying means comprises:
-
means for forming signals corresponding to
space="preserve" listing-type="equation">n'"'"'(i)=n(i)+m(i)-M.sub.tot /Kwhere each n'"'"'(i) is the modified vector quantized feature occupancy signal, each n(i) is the feature occupancy signal of the codebook ith feature vector signal, each m(i) is the occupancy signal for a corresponding one of the input feature vector signals classified as the ith codebook feature vector signal, Mtot is the total number of input feature vector signals, and K is the total number of the codebook quantized feature vector signals.
-
- 7. An arrangement for modifying a codebook of vector quantized feature vector signals according to claim 2 wherein said modifying means comprises forming a signal corresponding to
- space="preserve" listing-type="equation">q'"'"'(i)=(λ
)q(i)+(1-λ
)p(i)/m(i)
for each value of i, where each p(i) is the weighted input feature signal classified as the corresponding one of the vector quantized feature signals q(i) and each m(i) is the occupancy signal for said input feature signal, m(i)<
<
n(i), and λ
is a relative weighting factor which attributes substantially greater significance to n(i) than to m(i). - space="preserve" listing-type="equation">q'"'"'(i)=(λ
-
-
8. An arrangement for modifying a codebook of quantized feature vector signals according to claim 7 further comprising:
means for modifying said vector quantized feature occupancy signals responsive to said vector quantized feature occupancy signals and said input feature signal occupancy signals.
-
9. An arrangement for modifying a codebook of vector quantized feature signals according to claim 8 wherein said vector quantized feature occupancy signal modifying means comprises:
-
means for forming signals corresponding to
space="preserve" listing-type="equation">n'"'"'(i)=λ
n(i)+(1-λ
)m(i)for each value of is.
-
-
10. An arrangement for modifying a codebook of vector quantized feature signals according to claim 1, 3, 4, 5, 6, 7, 8 or 2 wherein each feature signal is a speech feature signal and each input pattern is a speech signal.
-
11. In a signal processing arrangement having a stored codebook of a set of vector quantized speech feature signals q(i) i=1, 2, . . . , K, where K is the number of vector quantized feature signals in said set, and occupancy signals n(i) each corresponding to one of said vector quantized feature signals, a method for modifying the codebook of vector quantized feature signals for speaker verification or adaptation to the voice of a particular speaker comprising the steps of:
-
receiving an input pattern; analyzing said input pattern to generate a set of input feature vector signals v(t), t=1, 2, . . . , M, where M is the number of vector-quantized feature signals in an input pattern; classifying each input feature vector signal v(t) as one of said set of vector quantized feature signals q(i) responsive to said input feature signals and said set of vector quantized feature signals; generating an occupancy signal m(i) corresponding to the count input feature vector signals classified as respective ones of said vector quantized feature signals q(i) responsive to the classifying step; and modifying said set of vector quantized feature vector signals, q(i), based on said classified input feature vector signals, said input feature occupancy signals, said vector quantized feature signals and said occupancy signals of said vector quantized feature signals. - View Dependent Claims (12, 13, 14, 15, 16, 17, 18, 19, 20)
-
15. In a signal processing arrangement having a stored codebook of a set of vector quantized feature signals q(i) i=1, 2, . . . , K and a set of occupancy signals n(i) each corresponding to one of said vector quantized feature signals, a method for modifying the codebook of vector quantized feature signals according to claim 14 further comprising the step of:
modifying each of said vector quantized feature occupancy signals responsive to each of said vector quantized feature occupancy signals and each of said input feature signal occupancy signals.
-
16. In a signal processing arrangement having a stored codebook of a set of vector quantized feature signals q(i) i=1, 2. . . , K and a set of occupancy signals n(i) each corresponding to one of said vector quantized feature signals, a method for modifying the codebook of vector quantized feature signals according to claim 15 wherein said vector quantized feature occupancy signal modifying step comprises:
-
forming a set of signals corresponding to
space="preserve" listing-type="equation">n'"'"'(i)=n(i)+m(i)-M.sub.tot /Kwhere n'"'"'(i) is the the set of modified vector quantized feature occupancy signals, n(i) is the feature occupancy signals of the codebook ith feature vector signals, m(i) is the set of occupancy signals for input pattern signals classified as respective ith codebook feature, Mtot is the total number of input feature signals, and K is the total number of the codebook quantized feature signals.
-
- 17. In a signal processing arrangement having a stored codebook of a set of vector quantized feature signals q(i) i=1, 2, . . . , K and a set of occupancy signals n(i) corresponding each to one of said vector quantized feature signals, a method for modifying the codebook of vector quantized feature signals according to claim 12 wherein said generating comprises forming a signal corresponding to
- space="preserve" listing-type="equation">q'"'"'(i)=(λ
)q(i)+(1-λ
)p(i)/m(i)
for each value of i, where each p(i) is the weighted input feature signal classified as one of the vector quantized feature signals q(i), each m(i) is the occupancy signal for said input feature signal, m(i)<
<
n(i), and λ
a relative weighting factor which attributes substantially greater significance to n(i) than to m(i). - space="preserve" listing-type="equation">q'"'"'(i)=(λ
-
-
18. In a signal processing arrangement having a stored codebook of a set of vector quantized feature signals q(i) i=1, 2, . . . , K and a set of occupancy signals n(i) each corresponding to one of said vector quantized feature signals, a method for modifying the codebook of vector quantized feature signals according to claim 17 further comprising the step of:
modifying each said vector quantized feature occupancy signal responsive to each said vector quantized feature occupancy signal and each said input feature signal occupancy signal.
-
19. In a signal processing arrangement having a stored codebook of a set of vector quantized feature signals q(i) i=1, 2, . . . , K and a set of occupancy signals n(i) each corresponding to one of said vector quantized feature signals, a method for modifying the codebook of vector quantized feature signals according to claim 18 wherein said vector quantized feature occupancy signal modifying step comprises:
-
forming a set of signals corresponding to
space="preserve" listing-type="equation">n'"'"'(i)=λ
n(i)+(1-λ
)m(i)where n'"'"'(i) is the set of modified vector quantized feature occupancy signals, n(i) is the set of feature occupancy signals of the codebook ith feature signals, m(i) is the set of occupancy signals for input feature signals classified as the ith codebook feature signals, and λ
is n(i)/(n(i)+m(i)), for each value of i.
-
-
20. An arrangement for modifying a codebook of vector quantized feature signals according to claims 11, 13, 14, 15, 16, 17, 18, 19 or 12 wherein each feature signal is a speech feature signal and each input pattern is a speech signal.
-
21. In a signal processing arrangement having a stored codebook of a set of vector quantized feature signals q(i) i=1, 2, . . . K, where K is the number of vector quantized feature signals in said set, and occupancy signals n(i) each corresponding to one of said vector quantized feature signals, a method for modifying the codebook of vector quantized feature signals, the method comprising the steps of:
-
receiving an input pattern; analyzing said input pattern to generate a set of input feature vector signals v(t), t=1, 2, . . . , M, where M is the number of vector-quantized feature signals in an input pattern; classifying each input feature vector signal v(t) as one of said set of vector quantized feature signals q(i) responsive to said input feature signals and said set of vector quantized feature signals; generating an occupancy signal m(i) corresponding to the count input feature vector signals classified as respective ones of said vector quantized feature signals q(i) responsive to the classifying step; and modifying said set of vector quantized feature vector signals, q(i), based on said classified input feature vector signals, said input feature occupancy signals, said vector quantized feature signals and said occupancy signals of said vector quantized feature signals.
-
Specification