US20080071538A1 - Speaker verification method - Google Patents

Abstract

A speaker verification method consist of the following steps: (1) generating a code book (42) covering a number of speakers having a number of training utterances for each of the speakers; (2) receiving a number of test utterances (44) from a speaker; (3) comparing (46) each of the test utterances to each of the training utterances for the speaker to form a number of decisions, one decision for each of the number of test utterances; (4) weighting each of the decisions (48) to form a number of weighted decisions; and (5) combining (50) the plurality of weighted decision to form a verification decision (52).

Description

FIELD OF THE INVENTION
• The present invention is related to the field of speech recognition systems and more particularly to a speaker verification method.
BACKGROUND OF THE INVENTION
• Speech recognition and speaker verification use similar analysis tools to achieve its goals. An input utterance is first processed to determine its essential characteristics. Typically, input utterances are converted to cepstrum coefficients. A cepstrum is an inverse Fourier transform of the log power spectrum. In a training phase the cepstrum coefficients are saved to form code books for specific utterances. For instance, a code book might have codes for the numeral zero through nine. In speech recognition, an input utterance is compared to the codes (training utterances) in the code book to determine which is most similar. In speech recognition the code is a generalized representation of many people's way of forming an utterance (e.g., “zero”). In speaker verification the codes represent the individual characteristics of the speaker and the verification system tries determine if a person's code is more similar to an input utterance, than an impostor code. As a result the codes in a speaker verification system emphasis individual characteristics, while in a speech recognition system the codes generalize over many individual speakers. Speaker verification has potential applications in a number of voice activated systems, such as banking over the telephone. Unfortunately, present speaker verification systems have not proven reliable enough for these applications.
• Thus there exists a need for an improved speaker verification system.
SUMMARY OF THE INVENTION
• An improved speaker verification method consist of the following steps: (1) generating a code book covering a number of speakers having a number of training utterances for each of the speakers; (2) receiving a number of test utterances from a speaker; (3) comparing each of the test utterances to each of the training utterances for the speaker to form a number of decisions, one decision for each of the number of test utterances; (4) weighting each of the decisions to form a number of weighted decisions; and (5) combining the plurality of weighted decisions to form a verification decision.
BRIEF DESCRIPTION OF THE DRAWINGS
• FIG. 1 is a block diagram of an embodiment of a speaker verification system;
• FIG. 2 is a flow chart of an embodiment of the steps used to form a speaker verification decision;
• FIG. 3 is a flow chart of an embodiment of the steps used to form a code book for a speaker verification decision;
• FIG. 4 is a flow chart of an embodiment of the steps used to form a speaker verification decision;
• FIG. 5 is a schematic diagram of a dial-up service that incorporates a speaker verification method;
• FIG. 6 is a flow chart of an embodiment of the steps used in a dial-up service; and
• FIG. 7 is a flow chart of an embodiment of the steps used in a dial-up service.
DETAILED DESCRIPTION OF THE DRAWINGS
• Several improvements in speaker verification methods are described and then a dial-up service that can incorporate these improvements is explained.FIG. 1 is a block diagram of an embodiment of aspeaker verification system10. It is important to note that the speaker verification system can be physically implemented in a number of ways. For instance, the system can be implemented as software in a general purpose computer connected to a microphone; or the system can be implemented as firmware in a general purpose microprocessor connected to memory and a microphone; or the system can be implemented using a Digital Signal Processor (DSP), a controller, a memory, and a microphone controlled by the appropriate software. Note that since the process can be performed using software in a computer, then a computer readable storage medium containing computer readable instructions can be used to implement the speaker verification method. These various system architectures are apparent to those skilled in the art and the particular system architecture selected will depend on the application.
• Amicrophone12 receives an input speech and converts the sound waves to an electrical signal. Afeature extractor14 analyzes the electrical signal and extracts key features of the speech. For instance, the feature extractor first digitizes the electrical signal. A cepstrum of the digitized signal is then performed to determine the cepstrum coefficients. In another embodiment, a linear predictive analysis is used to find the linear predictive coding (LPC) coefficients. Other feature extraction techniques are also possible.
• Aswitch16 is shown attached to thefeature extractor14. Thisswitch16 represents that a different path is used in the training phase than in the verification phase. In the training phase the cepstrum coefficients are analyzed by acode book generator18. The output of thecode book generator18 is stored in thecode book20. In one embodiment, thecode book generator18 compares samples of the same utterance from the same speaker to form a generalized representation of the utterance for that person. This generalized representation is a training utterance in the code book. The training utterance represents the generalized cepstrum coefficients of a user speaking the number “one” as an example. A training utterance could also be a part of speech, a phoneme, or a number like “twenty one” or any other segment of speech. In addition to the registered users samples, utterances are taken from a group of non-users. These utterances are used to form a composite that represents an impostor code having a plurality of impostor utterances.
• In one embodiment, thecode book generator18 determines whether the speaker (users and non-users) is male or female. The male training utterances (male group) are aggregated to determining a male variance vector. The female training utterances (female group) are aggregated to determine a female variance vector. These gender specific variance vectors will be used when calculating a weighted Euclidean distance (measure of closeness) in the verification phase.
• In the verification phase theswitch16 connects thefeature extractor14 to thecomparator22. Thecomparator22 performs a mathematical analysis of the closeness between a test utterance from a speaker with a training utterance stored in thecode book20 and between the test utterance and an impostor utterance. In one embodiment, a test utterance such as a spoken “one” is compared with the “one” training utterance and the “one” impostor utterance in thecode book20. Thecomparator22 determines a measure of closeness between the “one” training utterance, the “one” test utterance and the “one” impostor utterance. When the test utterance is closer to the training utterance than the impostor utterance, the speaker is verified as the true speaker. Otherwise the speaker is determined to be an impostor. In one embodiment, the measure of closeness is a modified weighted Euclidean distance. The modification in one embodiment involves using a generalized variance vector instead of an individual variance vector for each of the registered users. In another embodiment, a male variance vector is used for male speakers and a female variance vector is used for a female speaker.
• A decision weighting and combiningsystem24 uses the measure of closeness to determine if the test utterance is closest to the training utterance or the impostor utterance. When the test utterance is closer to the training utterance than impostor utterance, a verified decision is made. When the test utterance is not closer to the training utterance than the impostor utterance, an un-verified decision is made. These are preliminary decisions. Usually, the speaker is required to speak several utterances (e.g., “one”, “three”, “five”, “twenty one”). A decision is made for each of these test utterances. Each of the plurality of decisions is weighted and combined to form the verification decision.
• The decisions are weighted because not all utterances provide equal reliability. For instance, “one” could provide a much more reliable decision than “eight”. As a result, a more accurate verification decision can be formed by first weighting the decisions based on the underlying utterance. Two weighting methods can be used. One weighting method uses a historical approach. Sample utterances are compared to the training utterances to determine a probability of false alarm P_FA(speaker is not impostor but the decision is impostor) and a probability of miss P_M(speaker is impostor but the decision is true speaker). The P_FAand P_Mare probability of errors. These probability of errors are used to weight each decision. In one embodiment the weighting factors (weight) are described by the equation below:$a_i=\log \frac{1-P_{\mathrm{Mi}}}{P_{\mathrm{FAi}}}\mathrm{Decision}\mathrm{is}\mathrm{Verified}\left(\mathrm{True}\mathrm{Speaker}\right)$$a_i=\log \frac{P_{\mathrm{Mi}}}{1-P_{\mathrm{FAi}}}\mathrm{Decision}\mathrm{is}\mathrm{Not}\mathrm{Verified}\left(\mathrm{Imposter}\right)$
• When the sum of the weighted decisions is greater than zero then the verification decision is a true speaker. Otherwise the verification decision is an impostor.
• The other method of weighting the decisions is based on an immediate evaluation of the quality of the decision. In one embodiment, this is calculated by using a Chi-Squared detector. The decisions are then weighted on the confidence determined by the Chi-Squared detector. In another embodiment, a large sample approximation is used. Thus if the test statistics are t, find b such that c²(b)=t. Then a decision is an impostor if it exceeds the 1-a quantile of the c²distribution.
• One weighting scheme is shown below:
  1.5, if b>c_accept
  1.0, if 1−a≦b≦c_accept
  −1.0, ifc_reject≦b≦1−a
  −1.25, if b<c_reject
• When the sum of the weighted decisions is greater than zero, then the verification decision is a true speaker. When the sum of the weighted decision is less than or equal to zero, the decision is an impostor.
• In another embodiment, thefeature extractor14 segments the speech signal into voiced sounds and unvoiced sounds. Voiced sounds generally include vowels, while most other sounds are unvoiced. The unvoiced sounds are discarded before the cepstrum coefficients are calculated in both the training phase and the verification phase.
• These techniques of weighting the decisions, using gender dependent cepstrums and only using voiced sounds can be combined or used separately in a speaker verification system.
• FIG. 2 is a flow chart of an embodiment of the steps used to form a speaker verification decision. The process starts, atstep40, by generating a code book atstep42. The code book has a plurality of training utterances for each the plurality of speakers (registered users, plurality of people) and a plurality of impostor utterances. The training utterances in one embodiment are the cepstrum coefficients for a particular user speaking a particular utterance (e.g. “one). The training utterances are generated by a user speaking the utterances. The cepstrum coefficients of each of the utterances are determined to form the training utterances. In one embodiment a speaker is asked to repeat the utterance and a generalization of the two utterances is saved as the training utterance. In another embodiment both utterances are saved as training utterances.
• In one embodiment, a data base of male speakers is used to determine a male variance vector and a data base of female speakers is used to determine a female variance vector. In another embodiment, the data bases of male and female speakers are used to form a male impostor code book and a female impostor code book. The gender specific variance vectors are stored in the code book. Atstep44, a plurality of test utterances (input set of utterances) from a speaker are received. In one embodiment the cepstrum coefficients of the test utterances are calculated. Each of the plurality of test utterances are compared to the plurality of training utterances for the speaker atstep46. Based on the comparison, a plurality of decision are formed, one for each of the plurality of training utterances. In one embodiment, the comparison is determined by a Euclidean weighted distance between the test utterance and the training utterance and between the test utterance and an impostor utterance. In another embodiment, the Euclidean weighted distance is calculated with the male variance vector if the speaker is a male or the female variance vector if the speaker is a female. Each of the plurality of decisions are weighted to form a plurality of weighted decisions atstep48. The weighting can be based on historical error rates for the utterance or based on a confidence level (confidence measure) of the decision for the utterance. The plurality of weighted decisions are combined atstep50. In one embodiment the step of combining involves summing the weighted decisions. A verification decision is then made based on the combined weighted decisions atstep52, ending the process atstep54. In one embodiment if the sum is greater than zero, the verification decision is the speaker is a true speaker, otherwise the speaker is an impostor.
• FIG. 3 is a flow chart of an embodiment of the steps used to form a code book for a speaker verification decision. The process starts, atstep70, by receiving an input utterance atstep72. In one embodiment, the input utterances are then segmented into a voiced sounds and an unvoiced sounds atstep74. The cepstrum coefficients are then calculated using the voiced sounds atstep76. The coefficients are stored as a training utterance. For the speaker atstep78. The process then returns to step72 for the next input utterance, until all the training utterances have been stored in the code book.
• FIG. 4 is a flow chart of an embodiment of the steps used to form a speaker verification decision. The process starts, atstep100, by receiving input utterances atstep102. Next, it is determined if the speaker is male or female atstep104. In a speaker verification application, the speaker purports to be someone in particular. If the person purports to be someone that is a male, then the speaker is assumed to be male even if the speaker is a female. The input utterances are then segmented into a voiced sounds and an unvoiced sounds atstep106. Features (e.g., cepstrum coefficients) are extracted from the voiced sounds to form the test utterances, atstep108. Atstep110, the weighted Euclidean distance (WED) is calculated using either a generalized male variance vector if the purported speaker is a male. When the purported speaker is a female, the female variance vector is used. The WED is calculated between the test utterance and the training utterance for the speaker and the test utterance and the male (or female if appropriate) impostor utterance. A decision is formed for each test utterance based on the WED atstep112. The decisions are then weighted based on a confidence level (measure of confidence) determined using a Chi-squared detector atstep114. The weighted decisions are summed atstep116. A verification decision is made based on the sum of the weighted decisions atstep118.
• Using the speaker verification decisions discussed above results in an improved speaker verification system, that is more reliable than present techniques.
• A dial-up service that uses a speaker verification method as described above is shown inFIG. 5. The dial-up service is shown as a banking service. A user dials a service number on theirtelephone150. The public switched telephone network (PSTN)152 then connects the user'sphone150 with a dial-upservice computer154 at abank156. The dial-up service need not be located within a bank. The service will be explained in conjunction with the flow chart shown inFIG. 6. The process starts, atstep170, by dialing a service number (communication service address, number) atstep172. The user (requester) is then prompted by thecomputer154 to speak a plurality of digits (access code, plurality of numbers, access number) to form a first utterance atstep174. The digits are recognized using speaker independent voice recognition at,step176. When the user has used the dial-up service previously, verifying the user based on the first utterance atstep178. When the user is verified as a true speaker atstep178, allowing access to the dial-up service at step180. When the user cannot be verified, requesting the user input a personal identification number (PIN) atstep182. The PIN can be entered by the user either by speaking the PIN or by enter the PIN on a keypad. Atstep184 it is determined if the PIN is valid. When the PIN is not valid, the user is denied access atstep186. When the PIN is valid the user is allowed access to the service at step180. Using the above method the dial-up service uses a speaker verification system as a PIN option, but does not deny access to the user if it cannot verify the user.
• FIG. 7 is a flow chart of another embodiment of the steps used in a dial-up service. The process starts,step200, by the user speaking an access code to form a plurality of utterances atstep202. Atstep204 it is determined if the user has previously accessed the service. When the user has previously used the service, the speaker verification system attempts to verify the user (identity) atstep206. When the speaker verification system can verify the user, the user is allowed access to the system atstep208. When the system cannot verify the user, a PIN is requested atstep210. Note the user can either speak the PIN or enter the PIN on a keypad. Atstep212 it is determined if the PIN is valid. When the PIN is not valid the user is denied access atstep214. When the PIN is valid, the user is allowed access atstep208.
• When the user has not previously accessed the communication service atstep204, the user is requested to enter a PIN atstep216. Atstep218 it is determined if the PIN is valid atstep218. When the PIN is not valid, denying access to the service atstep220. When the PIN is valid the user is asked to speak the access code a second time to form a second utterance (plurality of second utterances) atstep222. The similarity between the first utterance (step202) and the second utterance is compared to a threshold atstep224. In one embodiment the similarity is calculated using a weighted Euclidean distance. When the similarity is less than or equal to the threshold, the user is asked to speak the access code again atstep222. In this case the second and third utterances would be compared for the required similarity. In practice, the user would not be required to repeat the access code atstep222 more than once or twice and the system would then allow the user access. When the similarity is greater than the threshold, storing a combination of the two utterances as atstep226. In another embodiment both utterances are stored as reference utterances. Next access to the service is allowed atstep208. The reference utterance (plurality of reference utterances, reference voiced sounds) is used to verify the user the next time they access the service. Note that the speaker verification part of the access to the dial-up service in one embodiment uses all the techniques discussed for a verification process. In another embodiment the verification process only uses one of the speaker verification techniques. Finally, in another embodiment the access number has a predetermined digit that is selected from a first set of digits (predefined set of digits) if the user is a male. When the user is a female, the predetermined digit is selected from a second set of digits. This allows the system to determine if the user is suppose to be a male or a female. Based on this information, the male variance vector or female variance vector is used in the speaker verification process.
• Thus there has been described an improved speaker verification method and a service that takes advantage of the speaker verification method. While the invention has been described in conjunction with specific embodiments thereof, it is evident that many alterations, modifications, and variations will be apparent to those skilled in the art in light of the foregoing description. Accordingly, it is intended to embrace all such alterations, modifications, and variations in the appended claims.

Claims (21)

1-21. (canceled)

22. A speaker verification method comprising:

comparing a test utterance corresponding to a speaker's input to a training utterance corresponding to the speaker;

comparing the test utterance to an imposter utterance; and

verifying the speaker as a true speaker in response to the test utterance being more similar to the training utterance than to the imposter utterance.

23. A method as defined inclaim 22, wherein the comparison between the test utterance and the training utterance and the comparison between the test utterance and the imposter utterance are determined using weighted Euclidean distances.

24. A method as defined inclaim 22, further comprising:

generating a preliminary verification decision indicative of whether the speaker is the true speaker based on whether the test utterance is more similar to the training utterance than to the imposter utterance;

weighting the preliminary verification decision based on at least one of a first probability value indicative of a probability of falsely determining that the true speaker is an imposter or a second probability value indicative of a probability of failing to detect an imposter speaker; and

verifying the speaker based on the weighted preliminary verification decision.

25. A method as defined inclaim 24, wherein the first and second probability values correspond to a first utterance different from a second utterance corresponding to third and fourth probability values different from the first and second probability values.

26. A method as defined inclaim 22, wherein verifying the speaker as the true speaker in response to the test utterance being more similar to the training utterance than to the imposter utterance is based on a historical error rate associated with an underlying phoneme of the input utterance.

27. A method as defined inclaim 22, further comprising generating the test utterance by identifying features of an input utterance from the speaker and generating a digital representation of the input utterance based on the identified features.

28. A machine accessible medium having instructions stored thereon that, when executed, cause a machine to:

compare a test utterance corresponding to a speaker's input to a training utterance corresponding to the speaker;

compare the test utterance to an imposter utterance; and

verify the speaker as a true speaker in response to the test utterance being more similar to the training utterance than to the imposter utterance.

29. A machine accessible medium as defined inclaim 28 having instructions stored thereon that, when executed, cause the machine to perform the comparison between the test utterance and the training utterance and the comparison between the test utterance and the imposter utterance using weighted Euclidean distances.

30. A machine accessible medium as defined inclaim 28 having instructions stored thereon that, when executed, cause the machine to:

generate a preliminary verification decision indicative of whether the speaker is the true speaker based on whether the test utterance is more similar to the training utterance than to the imposter utterance;

weight the preliminary verification decision based on at least one of a first probability value indicative of a probability of falsely determining that the true speaker is an imposter or a second probability value indicative of a probability of failing to detect an imposter speaker; and

verify the speaker based on the weighted preliminary verification decision.

31. A machine accessible medium as defined inclaim 30, wherein the first and second probability values correspond to a first utterance different from a second utterance corresponding to third and fourth probability values different from the first and second probability values.

32. A machine accessible medium as defined inclaim 28 having instructions stored thereon that, when executed, cause the machine to verify the speaker as the true speaker in response to the test utterance being more similar to the training utterance than to the imposter utterance based on a historical error rate associated with an underlying phoneme of the input utterance.

33. A machine accessible medium as defined inclaim 28 having instructions stored thereon that, when executed, cause the machine to generate the test utterance by identifying features of the input utterance and generating a digital representation of the input utterance based on the identified features.

34. A speaker verification method comprising:

receiving an input utterance from a speaker;

identifying voiced portions and unvoiced portions of the input utterance;

generating a test utterance based on the voiced portions but not the unvoiced portions of the input utterance; and

generating a verification decision indicative of whether the speaker is a true speaker based on the test utterance.

35. A method as defined inclaim 34, wherein the voiced portions correspond to at least vowel sounds of the input utterance.

36. A method as defined inclaim 34, further comprising generating the verification decision based on a historical error rate associated with an underlying phoneme of the input utterance.

37. A method as defined inclaim 34, further comprising generating the verification decision based on a preliminary verification decision that is weighted using at least one of a first probability value indicative of a probability of falsely determining that the true speaker is an imposter or a second probability value indicative of a probability of failing to detect an imposter speaker.

38. A machine accessible medium having instructions stored thereon that, when executed, cause a machine to:

receive an input utterance from a speaker;

identify voiced portions and unvoiced portions of the input utterance;

generate a test utterance based on the voiced portions but not the unvoiced portions of the input utterance; and

generate a verification decision indicative of whether the speaker is a true speaker based on the test utterance.

39. A machine accessible medium as defined inclaim 38, wherein the voiced portions correspond at least to vowel sounds of the input utterance.

40. A machine accessible medium as defined inclaim 38 having instructions stored thereon that, when executed, cause the machine to generate the verification decision based on a historical error rate associated with an underlying phoneme of the input utterance.

41. A machine accessible medium as defined inclaim 38 having instructions stored thereon that, when executed, cause the machine to generate the verification decision based on a preliminary verification decision that is weighted using at least one of a first probability value indicative of a probability of falsely determining that the true speaker is an imposter or a second probability value indicative of a probability of failing to detect an imposter speaker.

Images