US20080256613A1 - Voice print identification portal - Google Patents

Abstract

Systems and methods providing for secure voice print authentication over a network are disclosed herein. During an enrollment stage, a client's voice is recorded and characteristics of the recording are used to create and store a voice print. When an enrolled client seeks access to secure information over a network, a sample voice recording is created. The sample voice recording is compared to at least one voice print. If a match is found, the client is authenticated and granted access to secure information.

Systems and methods providing for a dual use voice analysis system are disclosed herein. Speech recognition is achieved by comparing characteristics of words spoken by a speaker to one or more templates of human language words. Speaker identification is achieved by comparing characteristics of a speaker's speech to one or more templates, or voice prints. The system is adapted to increase or decrease matching constraints depending on whether speaker identification or speaker recognition is desired.

Description

RELATED APPLICATIONS
• The present invention claims priority to U.S. Provisional Patent Application No. 60/894,627, entitled “VOICE PRINT IDENTIFICATION PORTAL,” filed Mar. 13, 2007 which is hereby incorporated by reference.
FIELD OF THE INVENTION
• The present invention relates generally to system access control based on user identification by biometric acquisition and speech signal processing for word recognition. More particularly, the present invention relates to combining voice based biometric identification for securing various computer related devices and speech recognition for device control and automated entry of information.
BACKGROUND
• The field of processing voice signals for use within a computerized device has traditionally been split into two distinct fields, speaker identification and speech recognition. These two fields have historically required separate and uniquely designed and configured systems. These systems are often provided by different vendors
• Speech recognition involves recognizing a human language word spoken by a speaker. In one example, speech recognition is utilized for computerized dictation, where a user speaks into a microphone and her words are recognized and entered into a document. Another example of speech recognition is controlling personal electronics, such as a cellular telephone or car stereo, through the use of verbal commands. Other applications for speech recognition include: command recognition, dictation, interactive voice response systems, automotive speech recognition, medical transcription, pronunciation teaching, automatic translation, and hands-free computing. Speech recognition is typically achieved through comparison characteristic qualities of spoken words, phrases, or sentences to one or more templates. A variety of algorithms are known in the art that allow qualification and/or comparison of speech to templates. These algorithms include: hidden Markov models, neural network-based systems, dynamic time warping based systems, frequency estimation, pattern matching algorithms, matrix representation, decision trees, and knowledge based systems. Some systems will employ a combination of these techniques to achieve higher accuracy rates.
• Speaker identification involves the process of identifying or verifying the identity of a specific person based on unique qualities of human speech. Human speech is often referred to as a biometric identification mechanism similar to finger prints or retinal scans. Like fingerprints and retinal scans, every individual has a unique voice print that can be analyzed and matched against known voice prints. Like other biometric identification mechanisms, voice prints can be utilized for verification or identification.
• Verification using a voice print is commonly referred to as voice authentication. Voice authentication is achieved in a similar manner to speech recognition: characteristic qualities of spoken words or phrases are compared to one or more templates. However, voice authentication is much more difficult to successfully achieve than speech recognition. First, speech recognition requires a less stringent match between the spoken word and a speech template. All that must be determined is what word was said, not who said that word based on a specific accent, pitch and tone. Second, speaker identification requires matching the speaker to a much larger number of possibilities, because one person must be identified out of many, not just what word they spoke. Whereas it may be acceptable to take up to several seconds to perform voice authentication, speech recognition must be done at a relatively fast pace in order for an interface to be reasonably useable.
• Traditionally, the use of speech for identification purposes versus speech for recognition purposes has been very segmented. While speech authentication requires complex and demanding comparisons, speech recognition requires real-time performance in order to meet user needs. Due to these differing requirements, existing systems (including computer hardware, software, or both) have been limited to performing one of these two functions.
• The use of speech to authenticate a user has a variety of advantages over other identification methods. First, like fingerprints or iris scans, every human being has an entirely unique speech pattern that can be quantifiably recognized using existing technology. Second, unlike fingerprints or iris scans, the input to a speaker identification system (the spoken word) may be different every time, even where the speaker is saying the same word. Therefore, unlike other methods of human authentication, speech authentication provides the additional advantage of an ability to prevent multiple uses of the same voice print.
• The rise of the computer age has drastically changed the manner in which people interact with each other in both business and personal settings. Along with the rise of the use of technology to conduct everyday life, security concerns with the use of computers have risen dramatically due to identity theft. Identity theft typically occurs where personal information such as bank accounts, social security numbers, passwords, identification numbers . . . etc., or corporate information is accessible when transferred over networks such as the internet, or when personal information or corporate information is entered into a user interface. For typical internet transactions such as consumer purchases, bank account transfers . . . etc, the transaction involves both a business side (back-end) and a customer side (front-end). The customer typically uses a computer, or handheld device such as a Smartphone or Personal Digital Assistant (PDA) to communicate during the transaction. Typically, communications during internet transactions are made very secure by using high security protocols such as Transport Layer Security (TSL) or Secure Socket Layer (SSL). However, when a customer enters in information (before it is transferred) at the front-end side of the transaction, the information is highly vulnerable to theft. In fact, in most cases of identity theft, personal information is stolen from the front-end side of the transaction. Therefore, a need exists to provide an efficient, more secure means of protecting the identity of one who wishes to interact in a secure environment over networks such as the internet. More specifically, a need exists to provide a secure transaction environment in which personal or corporate information is not communicated to the customer front-end in an accessible or repeatable format.
SUMMARY OF THE INVENTION
• The invention described herein seeks to remedy the issues discussed above by providing a system and method of voice authentication. In one embodiment, a method of securely authenticating a client seeking access to secure information or services available through a network is disclosed herein. In an embodiment, the method includes an enrollment process. The enrollment process may include receiving, at a server, an enrollment request and a voice recording. The process further includes processing, at the server, the voice recording to determine identifying characteristics of the client's voice, and creating a voice print identification of the client and storing the voice print identification.
• In an embodiment, the method also includes an authentication process. The authentication process includes receiving, at the server, a request for authentication of a client with an existing voice print. In one embodiment, the existing voice print was created according to the enrollment process discussed above. In one embodiment, the authentication process includes receiving a sample recording of the client's voice. In one embodiment, the process includes processing the sample recording. In one embodiment, the process includes comparing characteristics of the sample recording to at least one voice print identification. In one embodiment, the process includes determining, based at least in part on the comparing, that the client is authenticated. In one embodiment, the process includes communicating, over the network, an indication that the client is authenticated. In one embodiment, receiving, at the server, a sample recording of the client's voice is the only information received from the client that is used to determine that the client is authenticated.
• In another embodiment, a method of securely authenticating a client seeking access to secure information available through a network is described here. In an embodiment, the method includes an enrollment process. In an embodiment, the enrollment process includes sending, to a server, an enrollment request. In an embodiment, the enrollment process includes the voice recording of a client. In an embodiment, the enrollment process includes sending, to a server, the voice recording. In an embodiment, the enrollment process includes receiving, from the server, an indication that a voice print for the client has been created and stored based on the voice recording.
• In an embodiment, the method also includes an authentication process. In an embodiment, the authentication process includes sending, to the server, a request to authenticate the client. In an embodiment, the authentication process includes sending, to the server, a sample voice recording of the client. In an embodiment, the authentication process includes receiving, from the server, an indication that the client is authenticated. In an embodiment, the authentication process includes permitting the client access to secure information over the network based on the indication that the client is authenticated. In one embodiment sending, to the server, a sample voice recording of the client is the only information originating from the client that is used to authenticate the client.
• In an embodiment, a system for securely authenticating a client seeking access to secure information available through a network is described herein. In an embodiment, the system includes a back-end computer system adapted to manage and control access to secure information. In an embodiment, the system includes a front-end interface, adapted to provide the client with access to the back-end computer system. In an embodiment, the system includes a voice analysis computer system, adapted to verify a client's identity based on a voice sample. In an embodiment, the front-end interface is adapted to provide the client with the ability to record a client voice sample and communicate the client's voice sample to the voice analysis computer system. In an embodiment, the voice analysis computer system is adapted to compare the received client's voice sample to at least one voice print and authenticate the client based at least in part on the comparison. In an embodiment, the voice analysis computer system is adapted to communicate an indication of authentication. In an embodiment, the sample voice recording of the client is the only information originating from the client that is used to authenticate the client.
• In an embodiment, a method of operating a voice analysis system is described herein. In an embodiment, the method includes receiving, by a voice analysis system, at least one parameter indicating whether the system is to operate in a first mode or a second mode. In an embodiment, the method includes receiving, by the voice analysis system, a voice recording. In an embodiment, the method includes setting voice analysis constraints to a first level if the parameter indicates the first mode, or setting the voice analysis constraints to a second level if the parameter indicates the second mode. In an embodiment, the method includes comparing the voice recording to at least one template. In an embodiment, the comparison is based at least in part on the constraints. In an embodiment, the first mode indicates that the voice analysis system is to perform speaker identification. In an embodiment, the second mode indicates that the voice analysis system is to perform word recognition. In an embodiment, if the parameter indicates the first mode, an indication of authentication is provided. In an embodiment, if the parameter indicates the second mode, an indication of the textual value of the voice recording is provided.
BRIEF DESCRIPTION OF THE FIGURES
• The invention may be more completely understood in consideration of the following detailed description of various embodiments of the invention in connection with the accompanying drawings, in which:
• FIG. 1 illustrates generally a block diagram example of a typical transaction over a network.
• FIG. 2aandFIG. 2billustrate generally a flow chart example of one embodiment of voiceprint authentication.
• FIG. 3 illustrates generally an embodiment of the use of voice authentication for a client-business transaction over a network.
• FIG. 4 illustrates generally an alternative embodiment of the use of a voice authentication system.
• FIG. 5 illustrates generally one embodiment of an additional security feature.
• FIG. 6 illustrates generally one embodiment of an additional security feature.
• FIG. 7 illustrates generally one embodiment of a voice analysis system adapted to support both speech recognition and speaker identification.
• FIG. 8 illustrates generally a block diagram of one embodiment of an implementation of the system described herein.
• FIG. 9 illustrates generally one example of potential data stored by MySQL databases according to the subject matter described herein.
• FIG. 10 illustrates generally a flowchart example of an applet according to the subject matter described herein.
• FIG. 11 illustrates generally a flowchart example of a JSP application according to the subject matter described herein.
• FIG. 12 illustrates generally a flowchart example of a company administration JSP application according to the subject matter described herein.
• FIG. 13 illustrates generally a flowchart example of a web administration JSP application according to the subject matter described herein.
• While the invention is amenable to various modifications and alternative forms, specifics thereof have been shown by way of example in the drawings and will be described in detail. It should be understood, however, that the intention is not to limit the invention to the particular embodiments described. On the contrary, the intention is to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the invention as defined by the appended claims.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
• FIG. 1 illustrates generally a block diagram example of a typical transaction over a network. According to the example ofFIG. 1,client106 seeks to communicate with a business over a network such as the internet. To communicate,client106 uses a front-end interface101. Front-end interface101 may be any means with which a client may access content available over a network. A client may access a front-end interface101 through any commonly known network access device such as, but not limited to, a computer, a Smartphone, or a PDA.Client106, through front-end interface101, communicates with back-end computer system102. Back-end computer system102 may include a back-end server107.
• According to the example illustrated inFIG. 1, front-end interface101 communicates with back-end computer system102 throughconnection104. A typical transaction usually includes authentication ofclient106 by back-end computer system102. Often, authentication is achieved throughclient106 supplying some form of identification to back-end computer system102. Some examples of identification are pin numbers and associated passwords. Other examples include personal information such as social security numbers, addresses, telephone numbers, or client's106 mother's maiden name. Due to the need for personal devices discussed above to be able to connect with a large variety of sources,connection104 between front-end interface101 and back-end computer system102 is typically a less than secure connection. As a result of an insecure connection, the personal information transferred overconnection104 is vulnerable to theft.
• FIG. 2aandFIG. 2billustrate generally a flow chart example of one embodiment of voiceprint authentication according to the subject matter disclosed herein.FIG. 2aillustrates generally one embodiment of an account initialization process. At201, a client initiates an account with a provider of voiceprint authentication services (service provider). At202, the client is provided a means to record his/her voice and instructions for using those means. The client's voice may be recorded by any means known in the art, and in any format known in the art such as mp3 format, wav format, or a proprietary audio format. In various embodiments, it is to be understood that any digital audio format (e.g. wav, mp3 formats) is within the scope of the subject matter discussed herein. In an alternative embodiment, to ensure greater security, a proprietary audio format is used to record the client's voice. The recording is then transferred to the service provider. At203, the service provider analyzes the client's recording and extracts characteristics of the client's voice to create a voiceprint that represents the client's identity. At208, the voiceprint is stored by the services provider for later use in authenticating the client.
• FIG. 2billustrates generally a voiceprint authentication process according to the subject matter disclosed herein. It is assumed in this embodiment that prior to attempting voiceprint authentication of a client the client has gone through an account initialization process, such as the process discussed inFIG. 2a. At204, the client seeks voice authentication. The client may seek voice authentication for a number of reasons, including: internet website or telephone access to a bank or other commercial service provider, or in person authentication in a direct buyer/seller transaction. At205, the client is provided a means to record his/her voice. The client may be provided instructions to use a particular word or phrase, or the client may be allowed to speak any word or phrase for use in authenticating his identity. A recording of the client's voice is created. At206, the service provider compares the recording of the client's voice made atstep205 to stored voiceprints. In one embodiment, the service provider has access to only the client's voice recording, and that recording is compared to all the voiceprints the service provider has access to until a match is found. In another embodiment, the client's voice recording is provided to the service provider along with a client identification number. This embodiment is advantageous in that the service provider need only make a single comparison, thus decreasing the time and resources needed to authenticate a client. At207, if a match is found by the service provider, the client is authenticated.
• FIG. 3 illustrates generally one embodiment of the use of voice authentication system for a secure access transaction over a network according to the subject matter disclosed herein. In one embodiment,client301 seeks access to secure information or services. In order to do so,client301 uses front-end interface101 to accesswebpage302 that is available over a network such as the internet.Webpage302 is controlled by back-end computer system304. In typical transactions such as described herein,client301 is required to enter personal information such as an account number and/or password intowebpage302. As discussed with respect toFIG. 2, the connection between front-end interface101 andwebpage302 may be a non-secure connection. Therefore the opportunity for identity theft is present.
• According to the embodiment illustrated inFIG. 3, instead of password and username, a voiceprint is used to authenticate the client and allow him/her access to secure content on back-end computer system304. It is assumed for purposes of this discussion thatclient301 has previously initiated an account with the voice authentication service provider (service provider) as discussed inFIG. 2.
• Whenclient301visits webpage302, the client is offered the ability to, or required to, use voice authentication to access secure information. In various embodiments,client301 is provided with means to create a sample voice recording. In various embodiments, the client is provided an interface through the webpage to record his/her voice. The recording (and possibly a user id associated with the service provider as discussed in reference toFIG. 2) is communicated to voiceanalysis computer system303. Voiceanalysis computer system303 then compares the received recording to one or more stored voiceprints, and if a match is found, the client's identity is verified.
• In one embodiment, voiceanalysis computer system303 communicates, using a secure connection, with back-end computer system304 to determine whether theparticular client301 has permission to access particular content. In one embodiment, voiceanalysis computer system303 has access to a client security key (and possibly security keys allowing access to back-end computer system304 itself) that allows access to back-end computer system304. According to this embodiment, voiceanalysis computer system303 transmits the client security key to back-end computer system304. In response, back-end computer system304 may determine whetherclient301 should be granted access, and communicates (using a secure connection) authorization of access to voiceanalysis computer system303. Voiceanalysis computer system303 may then allow access to secure content throughwebpage302.
• In another embodiment, voiceanalysis computer system303 does not have access to a client security key to determine permission. Instead, voiceanalysis computer system303 attempts to verify the identity ofclient301, and, if successful, communicates success to back-end computer system304. According to this embodiment, back-end computer system304 determines whetherclient301 is to be granted permission to accesswebpage302, and back-end computer system304 itself communicates and allows access towebpage302.
• In another embodiment, voiceanalysis computer system303 verifies permission by reviewing client and business specific information stored on voiceanalysis computer system303. According to this embodiment, voiceanalysis computer system303 does not communicate security keys to back-end computer system304 and receive authorization from back-end computer system304. Instead, the entire authentication process is achieved in voiceanalysis computer system303. When a client's identity and permission are verified, authorized access is communicated towebpage302.
• The various embodiments of client authentication illustrated inFIG. 3 provide far greater security than previously known systems of client authentication. Because only the client's voice recording, and possibly service provider username, are entered into and/or communicated over accessible networks, it is nearly impossible for anyone to get access to the client's personal information. Identity verification through voice authentication provides a significant advantage in that it is nearly impossible to replicate a person's voice.
• In various other embodiments, client authentication illustrated inFIG. 3 may be used to authenticate access to content, information, or devices other thanwebpage302. Examples of such devices include cellular phones, computers, laptops, or a Personal Digital Assistant (PDA). In one such embodiment, client seeks access to a secure device through voice authentication. According to this embodiment, there may or may not be a back-end computer system304 that controls the device. In one embodiment, where back-end computer system304 does exist, the system functions identically to the above descriptions, however instead ofwebpage302 access, device access is controlled. In another embodiment, where no back-end computer system304 exists, the device is communicatively coupled to voiceanalysis computer system303 through a network. According to this embodiment,client303 is provided means to create a sample voice recording. The voice recording is communicated to voiceanalysis computer system303, and the client's301 voice may or may not be authenticated. Voiceanalysis computer system303 may have access to unlock, or otherwise provide access to, the device. If authentication is verified, voiceanalysis computer system303 communicates with the device and allowsclient301 access to the device.
• FIG. 4 illustrates generally an alternative embodiment of a voice authentication system to manage client-business transactions over a network according to the subject matter disclosed herein. The embodiment illustrated inFIG. 4 is nearly identical with the embodiment illustrated inFIG. 3, except voiceanalysis computer system403 is embedded within back-end computer system404.
• FIG. 5 illustrates generally one embodiment of an additional security feature according to the subject matter disclosed herein. The embodiment illustrated inFIG. 5 is nearly identical toFIG. 2b, except additional security steps are added to the process of voiceprint authentication. Similar toFIG. 2b, at501 a client seeks voice authentication to procure secured access. At502, the client is provided a means to record his/her voice, and a recording of the client's voice is created. At503, the service provider compares the recording of the client's voice made atstep502 to stored voiceprints. At504, the voice recording is compared with one or more existing voiceprints to determine if a match exists. Instead of authentication based primarily on the client's stored voiceprint alone, the embodiment illustrated inFIG. 5 includes the additional step of, at505, comparing the client's voice recording not only to the client's voiceprint, but also to one or more templates of recordings of the client's voice that were previously successful in the authentication process. This additional step ensures against fraudulent behavior in that it prevents someone from using a recording of a person's voice in order to access personal information. At505, if a voiceprint match is found, and the voice recording has been determined not to have been used previously, the client is authenticated. At506, the client's voice recording is stored for later comparison.
• FIG. 6 illustrates generally a flow chart of one embodiment of an additional security feature according to the subject matter presented herein. As previously discussed, audio data may be recorded and/or transmitted by any means known in the art, or by a proprietary format. In one embodiment, where a proprietary format is used, the audio data is further marked in order to provide additional security. Marking includes inserting, in the audio data, one or more indicators. These indicators are readable by systems adapted to utilize the proprietary data format. These indicators provide such systems with the ability to determine whether the audio data has been used previously, and whether the audio data has been used fraudulently. Therefore, the embodiment discussed above provides an additional layer of security to prevent the fraudulent use of audio data to access personal information. Turning now toFIG. 6, at601, a client's voice is recorded. At602, the client's voice is recorded in a proprietary audio format, or translated to a proprietary audio format, and an audio file is created. At603, markers are inserted into the audio file and the audio file is transmitted. At604, the audio data is received and processed. At605, the audio file and included markers are processed to determine if the audio file is the one sent, and whether or not the audio file has been determined fraudulent. At606, if the audio file is determined to be non-fraudulent, the audio file is used for purposes described herein.
• FIG. 7 illustrates generally one embodiment of aspeech analysis engine701 adapted to support both speaker identification and speech recognition. According to this embodiment, dual purposespeech analysis engine701 is adapted to accept as input aflag702 designating the desired function ofengine701.Flag702 indicates toengine701 whetherspeech recognition703 orspeaker identification704 is desired.Parameter constraints705 define the strictness with whichspeech processor708 determines a match according to speech characteristics. Ifspeaker identification704 is required,parameter constraints705 are set with strict requirements that must be met to find a positive match with a speaker's characteristics. In contrast, ifspeech recognition703 is desired, thenparameter constraints705 are set with much lower requirements that a positive match with the characteristics of a particular spoken word is found. Where speaker identification is desired,engine701 is adapted to return to the voice analysis computer system303 a positive or negative indication of whether the speaker was identified706. At707, if speech recognition is desired,engine701 is adapted to return a textual representation of spokenspeech707.
• FIG. 8 illustrates generally a block diagram of one embodiment of the implementation of avoice analysis system801 as disclosed herein. According to various embodiments,voice analysis system801 is implemented such that the system is capable of downloading and executing application software to front-end interface101. Such capabilities are advantageous to implementation ofsystem801 because they allow access and control of front-end interface101. For example,system801 may be capable of exercising control over microphone capabilities of front-end interface101. In one embodiment,system801 is implemented using a Java Virtual Machine environment. According to this embodiment,system101 includesJava Applet802.Applet802 is a program with the ability to download and execute software on front-end interface101.Applet802 controls much of the user interface requirements ofsystem801, such as microphone functionality.
• System801 may further includeJava JSP application803.Java JSP application803 is adapted to run on voiceanalysis computer system303. JSP application is further adapted to communicate withapplet802 to receive and transfer commands and information fromapplet802. In one embodiment,JSP application803 is adapted to receive a voice recording fromapplet802, and process that voice recording.System801 may further include one or more databases such as MySQL Database(s)804.JSP application803, among other applications, may be adapted to store and manage data inDatabases804.
• In some embodiments,system801 also includes Secure Web BasedAdministration Pages805. In various embodiments, administration pages805 provide an interface to create, modify, and configure client users.
• In some embodiments,system801 further includes Web Administration and Companyadministration JSP applications806. In various embodiments, Web Administration and CompanyAdministration JSP applications806 provide a web-based interface to configure companies, including companies access tosystem801.
• In one embodiment,applet802 is adapted to run on front-end interface101, whileJSP application803 is adapted to run on voiceanalysis computer system303. In an alternative embodiment, bothapplet802 andJSP application803 are adapted to run on front-end interface101. In yet another alternative embodiment,JSP application803 is adapted to run on back-end computer system304.
• FIG. 9 illustrates generally one example of potential data stored in MySQL Database(s)804. Detailed database schema SQL script source code is included as an appendix to this application.
• FIG. 10 illustrates generally a flow chart diagram of one embodiment of anapplet802 according to the subject matter disclosed herein. In one embodiment,applet802 is a Java Applet. In an alternative embodiment,applet802 is implemented as a Java Midlet instead of a Java Applet. A Java Midlet is preferred when front-end interface101 is a portable device or an embedded device. The term applet as used herein is intended to refer to either a Java Applet or a Java Midlet.Applet802 may be adapted to run on front-end interface101. At1002,applet802 is adapted to initialize on front-end interface101. Initializing may include determining the available sound recording/management hardware and software available to the front-end interface101. Initializing may further include receiving input parameters that define what actions Applet802 is to take. Input parameters may include: 1) whether enrollment, re-enrollment, authentication, or speech recognition capabilities are desired ofapplet802, and 2) identification information such as a user or company identification indicator. Initializing may further include providing a graphical user interface to a user such that the user may select input parameters forapplet802.
• At1003,applet802 is adapted to capture a client's voice. Voice capture may include: 1) providing a user interface to allow the client to record voice, 2) providing instructions to the client, 3) controlling front-end interface101 in order to record voice (including measuring background noise and setting detection thresholds), 4) verifying that the resultant recording meets requirements for further processing, and 5) preparing the recording for communication.
• In one embodiment, the voice recording is communicated using a TCP protocol. At1004, after the user's voice is recorded,applet802 sends the voice recording toJSP application803 for processing, and verifies that the communication was successful. In one embodiment,applet802 sends the voice recording over a secure connection such as an SSL connection. In one embodiment,JSP application803 runs on voiceanalysis computer system303.
• At1005, whenJSP application803 has completed processing the voice recording,applet802 processes return values fromJSP application803.Applet802 processes the return values based on what function was desired at1002. Also at1005,applet802 provides the user with a results display. In one embodiment, if authentication or enrollment were requested,applet802 provides the user with an indication that authentication was successful or unsuccessful. In another embodiment, where speech recognition was requested,applet802 provides the user with a textual indication of the words that were spoken. In a similar embodiment,applet802 provides the client with a verbal indication of words spoken by the client, orapplet802 may also act in response to words spoken by the client. Once the results have been provided to the user,applet802 returns to1002 and allows the client to re-enter parameters.
• FIG. 11 illustrates generally a flow chart diagram of one embodiment of aJava JSP application803 according to the subject matter disclosed herein. In one embodiment,Java JSP application803 is implemented as a Java Servlet. In another preferred embodiment,JSP application803 is run on voiceanalysis computer system303.
• At1102,JSP application803 awaits a request fromapplet802. When a request is received, JSP application processes the request. At1103,JSP application803, based on the request fromapplet802, determines what function is desired ofJSP application803.JSP application803 determines whetherapplet802 requested: enrollment of a new user, re-enrollment of an existing user, authentication of an enrolled user, or speech recognition.
• At1104, and1105, where enrollment of a new user or re-enrollment of an existing user is requested byapplet802, JSP application validates the user ID of the user, processes the voice recording, and updates an enrollment template and stores the template indatabases804. At1110, data is transferred back toapplet803.
• At1106 and1107, where authentication of an existing user is requested byapplet802, the user's user id is validated, the user's voice recording is processed, and the voice recording is compared to existing voice templates to determine whether the client is authenticated. If the client is authenticated, security tokens are prepared for transmission toapplet802. At1110, security tokens and other data are communicated toapplet802.
• At1109, where speech recognition is requested,JSP application803 is adapted to modify (lessen) voice recognition constraints such thatJSP application803 is only adapted to verify a particular word, not a particular client's voice. At1108, the voice recording is processed and compared to stored voice commands. If a match is found, an identification of a voice command is prepared for communication toapplet802. At1110, the identification of a voice command and other data are communicated toapplet802.
• FIG. 12 illustrates generally one embodiment of a CompanyAdministration JSP Application1201 as disclosed herein. CompanyAdministration JSP Application1201, provides an interface to create, modify, and configure client user data. The client user data consist of the following; UserID, CompanyName, FirstName, LastName, Password, Department, Position, Location, Email, Phone1, Phone2, Phone3, SecurityTokenID's, EnrollmentTemplateID's, and VoiceCommandID's.
• FIG. 13 illustrates generally one embodiment of a WebAdministration JSP Application1301. WebAdministration JSP Application1301 provides a web-based interface to configure companies, and their access to the Voice Print Portal Solution.
• In various embodiments, alternatives are provided for a client who does not have access to a front-end interface101 that is capable of recording voice. In one embodiment, a client is provided the ability to select a “Call In” button. When the “Call In” button has been selected, the client is provided an ordinary telephone number. The user may call the number in order to record his/her voice.
• In another embodiment, the client does not have any access to a front-end interface101 or the internet. According to this embodiment, a client is provided with the ability to operate the entire system through ordinary telephone service. The client may communicate with andrequest system801 functions through voice commands or though dialing numbers on a telephone keypad. In one embodiment, this telephone only system is implemented using telephony systems such as IPPC or IPPC express offered by Cisco Systems, Inc.
• Finally, while the present invention has been described with reference to certain embodiments, those skilled in the art should appreciate that they can readily use the disclosed conception and specific embodiments as a basis for designing or modifying other structures for carrying out the same purposes of the present invention without departing from the spirit and scope of the invention as defined by the appended claims.

Claims (18)

1. A method of securely authenticating a client seeking access to secure information available through a network, comprising:

receiving, at a server, an enrollment request;

receiving, at the server, a voice recording;

processing, at the server, the voice recording to determine identifying characteristics of the speaker's voice;

creating, based on the identifying characteristics, a voice print identification of the speaker;

storing, at the server, the voice print identification;

receiving, at the server, a request for authentication of a client with an existing voice print identification;

receiving, at the server, a sample recording of the client's voice;

processing, at the server, the sample recording of the client's voice;

comparing characteristics of the sample recording to at least one voice print identification;

determining, based at least in part on the comparing, that the client is authenticated; and

communicating, over the network, an indication that the client is authenticated.

2. The method ofclaim 1, wherein the communicating an indication that the client is authenticated comprises:

communicating only non-critical information.

3. The method ofclaim 1, wherein the communicating an indication that the client is authenticated comprises:

communicating only a positive or negative indication that the client is authenticated.

4. The method ofclaim 1, wherein receiving, at the server, a sample recording of the client's voice further comprises:

receiving an indication of the client's identity.

5. The method ofclaim 1, wherein determining, based at least in part on the comparing, that the client is authenticated further comprises:

comparing the sample recording of the client's voice to at least one stored indication of characteristics of sample recordings that were previously used to authenticate the client; and

if the sample recording has been previously used to authenticate, not providing an indication that the client is authenticated.

6. The method ofclaim 1, wherein receiving, at the server, a sample recording of the client's voice is the only information received from the client used to authenticate the client.

7. A method of securely authenticating a client who seeks access to secure information available through a network, comprising:

sending, to a server, an enrollment request;

recording the voice of a client;

sending, to a server, the voice recording;

receiving, from the server, an indication that a voice print for the client has been created and stored based on the voice recording;

sending, to the server, a request to authenticate the client;

sending, to the server, a sample voice recording of the client;

receiving, from the server, an indication that the client is authenticated; and

permitting the client access to secure information over the network based on the indication that the client is authenticated.

8. The method ofclaim 7, wherein the receiving an indication that the client is authenticated comprises:

receiving only non-critical information.

9. The method ofclaim 7, wherein the receiving an indication that the client is authenticated comprises:

receiving only a positive or negative indication that the client is authenticated.

10. The method ofclaim 7, wherein sending, to the server, a sample voice recording of the client further comprises:

sending an indication of the client's identity.

11. The method ofclaim 7, wherein sending, to the server, a sample voice recording of the client is the only information originating from the client that is used to authenticate the client.

12. A system for securely authenticating a client seeking access to secure information available through a network, comprising:

a back-end computer system adapted to manage and control access to secure information;

a front-end interface, adapted to provide the client with access to the back-end computer system;

a voice analysis computer system, adapted to verify a client's identity based on a voice sample;

wherein the front-end interface is adapted to provide the client with the ability to record a client voice sample and communicate the client's voice sample to the voice analysis computer system;

wherein the voice analysis computer system is adapted to compare the received client's voice sample to at least one voice print and authenticate the client based at least in part on the comparison; and

wherein the voice analysis computer system is adapted to communicates an indication of authentication.

13. The system ofclaim 12, wherein the voice analysis computer system is adapted to communicate the indication of authentication to the front-end interface, and wherein the front-end interface is adapted to allow the client access to secure information.

14. The system ofclaim 12, wherein the voice analysis computer system is adapted to communicate the indication of authentication to the back-end computer system, and wherein the back-end computer system is adapted to allow the client access to secure information.

15. The system ofclaim 12, wherein the voice analysis computer system is adapted to allow the client access to secure information.

16. The system ofclaim 12, wherein the front-end interface is adapted to provide the client with the ability to record a client voice sample and communicate the clients voice sample to the voice analysis computer system along with an indication of the client's identity.

17. The system ofclaim 12, wherein the sample voice recording of the client is the only information originating from the client that is used to authenticate the client.

18. A method of operating a voice analysis system, comprising:

receiving, by a voice analysis system, at least one parameter indicating whether the system is to operate in a first mode or a second mode;

receiving, by the voice analysis system, a voice recording;

setting voice analysis constraints to a first level if the parameter indicates the first mode, or setting the voice analysis constraints to a second level if the parameter indicates the second mode; and

comparing the voice recording to at least one template, wherein the comparison is based at least in part on the constraints, wherein the first mode indicates that the voice analysis system is to perform speaker identification, wherein the second mode indicates that the voice analysis system is to perform word recognition; and

wherein if the parameter indicates the first mode, providing an indication of authentication, and wherein the if the parameter indicates the second mode, providing an indication of the textual value of the voice recording.

Images