KR19990026429A - PCS user verification method using voice identification - Google Patents

Abstract

The present invention relates to a personal mobile communication service (PCS), and relates to a PCS user verification method for determining whether to provide an additional service through voice identification of a PCS user.

According to the present invention, when a user who wants to use the PCS supplementary service inputs an supplementary service code, the user inputs an arbitrary password requested by the supplementary service providing system by voice, and the user's voice according to the speech identification algorithm on the supplementary service providing system side. And analyzing the password to determine whether the subscriber is a registered subscriber and providing the supplementary service only to the legitimate subscriber.

The present invention can prevent damage caused by the leakage and theft of passwords assigned by a conventional series of numbers by verifying whether the subscriber is registered only by the user's voice, and can not remember the password and receive additional services. Can be solved.

Description

Method of verifying for PCS user's authentication using voice-based identification

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to personal communications services (hereinafter referred to as PCS), and to a PCS user verification method for determining whether to provide an additional service through voice identification of a PCS user who wants to receive an additional service.

In general, the PCS service is a mobile communication service that enables voice and data exchange between individuals regardless of time and place by assigning individual numbers to users. PCS is based on call service and provides services by various additional service providers such as home shopping, home banking and voice / fax message service.

If the PCS user wants to use an additional service such as home shopping or home banking, the PCS user selects a desired service type by using the keypad. The additional service provider requests a password assigned to the user, assigns a service use right only by checking the password input by the user.

That is, at present, only a password input through a PCS terminal when using a service is only a verification procedure for authenticating a user. The user verification method by inputting a password may cause a problem that a legitimate user cannot use the service because he or she cannot remember the password, and there is a serious problem that others can steal when the password is lost or leaked.

Accordingly, the present invention has been made to solve the above problems, the present invention provides a service only when the voice of the registered subscriber to the PCS user who wants an additional service to input a random password by voice. The purpose is to provide a PCS user verification method using voice identification.

The method of the present invention for achieving the above object, in the verification procedure for confirming the identity of the PCS user when the user wants to receive the additional service using the PCS terminal, the PCS user through the keypad of the terminal desired additional service Inputting a code, the additional service providing system generates a random password in order to perform a voice identification function, and sends a request message to the user to input the randomly generated password by voice; When the PCS user inputs the requested password as a voice, the additional service provider collects the user's voice and runs a voice identification algorithm to analyze the user's voice and password; The user's voice analyzed according to the voice identification algorithm is used to determine the identity of the registered subscriber, and if the subscriber is determined to be a registered subscriber, additional services are started, and if it is determined that the user is not a registered subscriber, the subscriber registered with the user is determined. Sending a guide message to indicate that the message does not match and to repeat the verification process; In order to repeat the verification process for the failed user, the additional service providing system repeats the same verification process by regenerating a random password, and if it is determined that the voice of the registered subscriber is not found even through N verification processes. Determining that the illegal user is characterized in that it comprises a step of disconnecting.

1 is a block diagram of a general speaker recognition system;

2 is a flowchart illustrating a verification method for PCS user authentication according to the present invention.

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

Since the size and shape of the pronunciation organs differ from person to person, and the pronunciation habits acquired during the growth process are different from each other, even if the same sentence is pronounced, the voice signal includes information about the speaker. Thus, voice can distinguish an individual's identity, such as a unique fingerprint or retinal pattern.

Verification or authentication of a speaker's identity using information extracted from the voice signal is called speaker recognition. Speaker identification (or speaker identification) is the identification of a speaker in a population by the speaker's input voice, and the speaker's demonstration of whether or not the input voice is a particular person claimed by the speaker who pronounced it. This is called speaker verification.

The configuration of a general speaker recognition system will be described with reference to FIG. 1 by dividing the speaker verification process and the speaker identification process.

The speaker verification process (speaker verification process) is to extract the feature parameters from the speech signal for each speaker in 10 to create a standard pattern, if the unknown voice is input in 11 to extract the feature parameters to create a test pattern, The pattern matching is performed by pattern matching of 10 standard patterns and 11 test patterns, and a threshold value is set for speaker verification at 13 and compared with the distance between 12 patterns for speaker verification. Decide A speaker verification decision of 13 determines 'approve' if the calculated distance of 12 is less than the threshold set for that speaker and 'decline' if the calculated distance of 12 is above the threshold set for that speaker.

In the speaker identification process, the speaker's voice signal is input to extract a feature parameter to generate a test pattern, and at 12, a pattern matching is performed between all speaker's standard patterns and 11's test pattern. The distance between the patterns is calculated, and if the distance between the patterns is 14 and 12 and the threshold is less than the threshold, the speaker having the minimum distance is determined as the speaker for the input voice. If the threshold is exceeded, the speaker not registered in the system is not. To judge.

Currently, the algorithms used for voice recognition or voice recognition generally include classical dynamic program algorithms, HMM (Hidden Markov Moddel) by probabilistic methods, and methods using artificial neural networks. Known.

The present invention is described in detail below with reference to a specific embodiment as a method to apply the speaker recognition system to provide a supplementary service only to the true user verified by analyzing the voice of the PCS user.

1 is a flowchart illustrating a PCS user verification method using voice identification according to the present invention.

When the PCS user inputs a desired supplementary service code through the keypad of the terminal (a), the supplementary service providing system generates an arbitrary password to perform a voice identification function (b), and inputs any generated password by voice. Send a request message to the user (c).

When the PCS user inputs the requested password by voice (d), the additional service provider collects the user's voice and runs a voice identification algorithm to analyze the user's voice and password (e).

It is determined whether the user's voice analyzed according to the voice identification algorithm is a registered subscriber stored in the database to confirm the identity (f), and if it is determined that the registered subscriber is an additional service, the additional service is started (f).

If it is not a registered subscriber, the user is notified that the registration does not match with the registered subscriber, and a guidance message is sent to repeat the verification process (g). And, in order to repeat the verification process for the failed user, the additional service providing system performs the same verification process such as regenerating a random password.

If it is determined through the N verification process that the voice of the registered subscriber is not (h), it is determined that the illegal user is released (i).

In (b) to (d), a random password is generated at every service request, and the user inputs it by voice. When using a unique password, the voice of the unique password may be recorded and other persons may steal it. This is to prevent this because it cannot be excluded. In addition, even if the registered subscriber does not remember their password, the service can be used only by their own voice.

As described above, the conventional verification process has problems such as the risk of leaking and stealing the password given to the PCS user, the hassle caused by the change of the password, and the failure to remember the password and not being provided with additional services. The present invention can solve the above-mentioned problems by verifying whether the subscriber is registered only by the voice of the user, and in particular, there is an effect of thoroughly excluding the illegal user.

Claims (1)

In the verification procedure for confirming the identity of the PCS user when the additional service is to be provided using the PCS terminal, (a) a PCS user entering a desired additional service code through a keypad of a terminal; (b) generating, at the supplementary service providing system side, an arbitrary password to perform a voice identification function; (c) the supplementary service providing system side sending a request message to the user for inputting any generated password by voice; (d) the PCS user voice inputting the required password; (e) collecting, by the additional service provider, a voice of the user and running a voice identification algorithm to analyze the voice and password of the user; (f) determining an identity by determining whether the voice of the user analyzed according to the voice identification algorithm matches one of the registered subscribers; (g) starting the additional service if it is determined that the subscriber is a registered subscriber (f); (h) If it is determined in (f) that it is not a registered subscriber, it is examined whether it has failed more than N times, and if less than N times, the user is not matched with the registered subscriber, and the verification process is repeated again. Sending a guide message and repeating steps (b) to (f) in order; (i) PCS user verification method using the voice identification, characterized in that comprises the step of determining that the user is not a voice of the registered subscriber (N) more than N times, the illegal user to disconnect the connection. .