KR100677669B1 - Apparatus and method for generating a pattern for behavior-based User Authentication - Google Patents

Abstract

Disclosed are a method and apparatus for generating a pattern for user authentication. The present invention provides a method for generating a user pattern representing a user's password typing pattern, which is used in a user behavior based user authentication system, the method comprising: (a) providing a regular signal to be a reference timing at the start of a user's password input; Rendering the timing queue to the user; (b) receiving a keystroke corresponding to the password, input according to the timing queue; And (c) generating the user pattern by calculating a timing vector for the keystrokes. According to the present invention, the quality of the pattern is increased and as a result the error rate of user authentication is reduced.

User authentication, behavioral, timing

Description

Apparatus and method for generating a pattern for behavior-based User Authentication

1 is a diagram illustrating an example of a typing pattern expressed by a timing vector.

2 is a view showing a conventional user behavior based user authentication device.

3 is a view showing a conventional pattern generation method.

4 is a view showing a pattern generation method according to an embodiment of the present invention.

5 is a diagram illustrating an example of a timing queue.

6A to 6H illustrate typing patterns according to various password input methods.

7 shows the quality of a pattern generated in the manner of FIGS. 6B-6H.

8A to 8H are diagrams showing the distance between the circular pattern, the typing pattern, and the impersonator pattern in FIG.

9A and 9B are diagrams comparing a conventional pattern with a pattern according to the present invention.

10 is a graph showing the results in Table 2. FIG.

11 is a view showing a pattern generating apparatus according to an embodiment of the present invention.

12 illustrates a user authentication device according to an embodiment of the present invention.

13 is a view showing a user authentication method according to an embodiment of the present invention.

The present invention relates to a pattern generating apparatus and method for behavior-based user authentication, and more particularly, to a pattern generation method for increasing the quality of a pattern by increasing the consistency of the user pattern and consequently reducing the error rate of user authentication. A device, and a user authentication method and device.

User authentication is to determine whether the user currently attempting to log on is a true user. Currently widely used user authentication method is user authentication through password input. User authentication through password entry is performed by comparing whether the password entered by the user and the true password previously entered into the device to log on match. However, user authentication based on the password comparison method cannot prevent illegal intrusion when an impersonator steals a user's password.

Two new methods have been introduced to solve this problem. One is to perform user authentication using the user's physical characteristics, and the other is to perform user authentication using the user's behavior pattern.

The first method is a biometrics method based on physical characteristics. A user who wants to log on by storing the user characteristic information generated by analyzing the user's physical characteristics, for example, the user's fingerprints and irises. This is done by obtaining the user characteristic information of and comparing the two. This method has a problem in that the cost for increasing the accuracy of the analysis device increases because the accuracy of the user authentication is determined according to the accuracy of the device analyzing the user characteristic information. In addition, the analysis and comparison process is relatively complicated, and the implementation cost itself is high.

The second method is a behavior-based biometric method that uses user characteristic information generated by analyzing user's behavior patterns, especially typing patterns. According to this method, first, a user's typing pattern is stored, and user authentication is performed by comparing whether a user's typing pattern input at the time of logon is identical to a pre-stored real user's typing pattern. According to this method, since the typing pattern is a result of the user's conscious behavior, the quality of the pattern can be increased according to the user's effort, the cost of implementation is low, and the implementation procedure is simple.

1 is a diagram illustrating an example of a typing pattern expressed by a timing vector.

The typing pattern is obtained from the user's keystrokes. Since keystrokes are characterized by how long a key is pressed and by the input time interval between the key, the typing pattern is typically represented by a timing vector. That is, the timing vector represents a typing time as a vector corresponding to a password unit that is typed by a user, that is, letters and numbers. Fig. 1 shows an example of a timing vector when the password is “password.” As shown in Fig. 1, the timing vector when the password is 8 characters becomes a 17-dimensional vector, for example (2, 4, 6,3,4,5,8,0,0,0,2,4,4,5,6,6,9).

2 is a diagram illustrating a conventional user behavior based user authentication apparatus.

The user authentication apparatus 200 includes a pattern generator 210 and a determiner 220. The pattern generator 210 generates the typing pattern 212 by analyzing the user's keystroke 202. The typing pattern 212 is represented by a timing vector as shown in FIG. 1. The determination unit 220 generates an authentication result value 222 based on the typing pattern 212, the circular pattern 214, and the threshold 216. Specifically, the determination unit calculates a difference value between the typing pattern 212 and the circular pattern 214, and generates an authentication result value 222 indicating that user authentication is successful only when the difference value is smaller than the predetermined threshold value 216. do. The prototype pattern 214 is a value previously input and stored as a typing pattern of a true user.

3 is a diagram illustrating a conventional typing pattern generation method.

In operation 310, the pattern generating apparatus displays a password window.

In step 320, the user types a password in the password window. That is, the pattern generating device receives the keystrokes.

In operation 330, the apparatus for generating a pattern generates a typing pattern by analyzing keystrokes and calculating a timing vector.

However, according to the pattern generation method of FIG. 3, the generated pattern is inconsistent. This is because the keystrokes entered by the user are not constant with each typing. Since the circular pattern, which is the basis of user authentication, is also generated by the pattern generation method of FIG. 3, this problem greatly affects the accuracy of user authentication. Therefore, in the implementation of the user authentication device based on user behavior, increasing the consistency of the pattern, that is, the quality of the pattern, has been the most important issue.

Accordingly, the present invention has been made to solve the above-described problem, and to provide a pattern generating apparatus and method for increasing the quality of a pattern in user behavior based user authentication, and a user authentication apparatus and method using the same.

The present invention for solving the above problems, in a method for generating a user pattern indicating a user's password typing pattern, which is used in a user behavior-based user authentication system, (a) reference timing at the start of a user's password input Rendering a timing queue to the user, the timing queue providing a regular signal to be; (b) receiving a keystroke corresponding to the password, input according to the timing queue; And (c) generating the user pattern by calculating a timing vector for the keystrokes.

Here, the timing queue is any one of an audio signal, a video signal, or a combination thereof.

In addition, the present invention, in the apparatus for generating a user pattern indicating the password typing pattern of the user, used in a user behavior based user authentication system, (a) a regular signal to be a reference timing at the start of password input of the user A timing queue rendering unit for rendering a timing queue to a user; (b) a timing vector calculator configured to generate the user pattern by calculating a timing vector for a keystroke of a user input according to the timing queue.

In addition, the present invention provides a user behavior based user authentication apparatus, comprising: (a) a timing queue rendering unit for rendering a timing queue for providing a regular signal to a user so that a reference signal is provided when a user inputs a password; a timing vector calculator configured to generate a user pattern representing a user's typing pattern by calculating a timing vector of an input keystroke based on the timing queue; And (c) a determination unit that determines whether the user is authenticated based on the user pattern, a training pattern representing a typing pattern of a true user, and a threshold value.

Here, the timing queue is the same as the timing pattern used when generating the training pattern.

In another aspect, the present invention provides a user behavior-based user authentication method comprising the steps of: (a) rendering a timing queue for providing a regular signal to a user so as to be a reference timing when a user inputs a password; (b) generating a user pattern representing a user's typing pattern by calculating a timing vector of an input keystroke based on the timing queue; And (c) determining whether to authenticate the user based on the user pattern, the training pattern, and the threshold value.

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

A user is a human entity to be authenticated and includes a true user and an impersonator.

The pattern or typing pattern is a pattern generated from a keystroke of a password by a true user or impersonator. In the present invention, the typing pattern may be represented by a timing vector.

The training pattern represents a typing pattern generated from keystrokes entered by a true user at password registration.

The circular pattern is the mean value of the training patterns.

The classification reference value is a value mathematically representing a relationship between a true pattern of a user and a pattern input during user authentication, and the classifier determines whether to authenticate by comparing the classification reference value with a threshold value.

The test pattern represents a typing pattern generated from keystrokes entered by a true user upon password authentication.

The impersonator pattern represents the typing pattern of the impersonator.

A password is a value that a user or impersonator enters into a user authentication device for user authentication.

The password element is a unit that constitutes a password and includes letters and numbers. The password element corresponds to one keystroke on the keyboard.

4 is a view showing a pattern generation method according to an embodiment of the present invention.

In step 410, a password window is displayed to the user.

In step 420, a timing queue is rendered to the user. The timing queue is an audio signal, a video signal, or a combination thereof that provides reference timing to facilitate user password entry. The timing queue is rendered to the user from when the password input window is displayed to the user so that the user can enter a password. The timing queue is rendered to the user as a regular and periodic signal. Users can enter passwords while listening to these periodic signals, thereby increasing the consistency of their typing patterns. For example, if putting a pause between the first and second characters is the user's typing pattern, the user can more easily determine when the pause is going by listening to a timing cue, which is a periodic signal.

In step 430, the user enters the password by keystrokes typing corresponding to his password according to the timing queue. The pattern generation device receives a keystroke.

In operation 440, the apparatus for generating a pattern generates a password pattern by calculating a timing vector for keystrokes. Keystrokes store timings and pressed keys each time a key on the keyboard corresponding to the key unit (consonants, vowels, alphabets, numbers, etc.) included in the password is pressed and represented in the form of a vector to express the timing vector. Calculate

5 is a diagram illustrating an example of a timing queue.

Timing queues allow users to be consistent in password entry by recognizing the time signature. In other words, the timing queue starts to be rendered when the user is prompted for a password, and provides the user with a rhythm or rhythm to the password input. Since the user sees or hears (or hears and sees) the timing queue being rendered, the user can easily remember the typing pattern of the password that the user previously entered when registering the password, thereby maintaining the consistency of the pattern.

The timing cue may be generated from an audio signal, a video signal, or a combination thereof. Figure 5 is a video clip consisting of a total of four still images is an example of a video timing queue, one cycle is generated each time the hammer hits, the user is looking at the third picture the hammer hits the nails This allows the user to have a rhythmic sense of periodicity, thereby maintaining consistency in the typing pattern for entering passwords.

In addition to the timing cue, audio signals such as clock second hand sounds are possible, and a combination of the two signals as described above is also possible.

The user authentication device determines whether the user is authenticated by comparing the previously stored circular pattern with the analyzed typing pattern when the user inputs the password. The quality of patterns used in user behavior-based user authentication devices is evaluated in three aspects.

The first is uniqueness. How unusual is the typing pattern in which the user enters a password? That is, it is a measure of how distinguished it is from others. The higher the specificity, the easier the user authentication device can distinguish between the typing pattern of another person and the typing pattern of a true user.

The specificity in the present invention is generated based on the difference between the distance between the imposter pattern and the circular pattern and the distance between the training pattern and the circular pattern. For example, the following equation (1).

[Equation 1]

Here, U is specificity, M is circular pattern, Z _k is impersonator pattern, N _z is input number of typing pattern of impersonator, X _i is training pattern, N _x is input number of training pattern.

The second is consistency, a measure of how consistent the user's pattern is. Even if the user is a true user, if the user's typing pattern is changed every time, the user authentication device cannot make a correct decision. Therefore, maintaining a consistent typing pattern of the user plays a large role in reducing the error rate of the user authentication device. In general, inconsistency is used as a value for numerical measurement.

In the present invention, the inconsistency is generated based on the difference between the distance between the test pattern and the circular pattern and the distance between the training pattern and the circular pattern. For example, it is as follows.

[Equation 2]

Here, I is an inconsistency, M is a circular pattern, Y _j is a test pattern, N _y is an input number of test patterns, X _i is a training pattern, and N _x is an input number of training patterns.

The third is discriminability, which is a measure of how distinguishing a user's pattern is from other users. Discrimination is a concept created by mathematically combining singularity and consistency, rather than independent meanings such as specificity or consistency.

In the present invention, the discrimination degree is generated based on the difference between the minimum distance between the impersonator pattern and the circular pattern and the maximum distance between the test pattern and the circular pattern. For example, it is as follows.

[Equation 3]

Here, D is a discrimination degree, M is a circular pattern, Z _k is a pseudonym pattern, and Y _j is a test pattern.

This mathematical definition of specificity, consistency and discrimination allows for a more accurate measurement of the pattern's quality.

The administrator of the user authentication device should select the threshold value as an appropriate value to reduce the error rate of the user authentication device. At this time, if the degree of discrimination can be increased, the range of an appropriate threshold value can be increased.

Hereinafter, the effect of the timing queue will be described using the quality of the pattern using the definitions of Equations 1 to 3.

6A through 6H illustrate typing patterns according to various password input methods.

The user may increase the specificity of the typing pattern by entering a password in accordance with artificial rhythm. 6A shows the user's typing pattern when there is no artificial rhythm, and FIGS. 6B through 6H show the typing pattern generated by the five artificial rhythms. The five artificial rhythms here are pose, musical rhythm, staccato, legato and slow tempo.

6B and 6C show typing patterns with pauses. The user can insert artificial time margins between the letters or numbers that make up a password. For example, you can type pa__ss__word when entering the password "password". Where "_" represents one pose. In general, the interval between characters generated when a user enters a password is 50 to 200 ms. Therefore, when a user artificially inserts a pose, the interval increases to 2000 ms, which increases the specificity of the user's typing pattern. The user can insert this artificial pose by tapping twice outside the keyboard. The user must remember where and for how long the pose was inserted. In general, the longer the pose, the specificity of the user's typing pattern increases (FIG. 6C).

6D illustrates a typing pattern using musical rhythm. A tune, melody, which the user knows may be used to increase the specificity of the user's typing pattern. 6D illustrates a typing pattern generated based on the music rhythm of the cheering song used by the Korean soccer cheering party in the 2002 World Cup. The music rhythm has its own advantages in that it provides specificity and is also easy for the user to remember.

6E and 6F are typing patterns using staccato. Staccato is a term for a short break in a piano, and a user may increase the specificity of a typing pattern by shortly typing and typing each character when a password is input. FIG. 6E shows a staccato of one unit length, and FIG. 6F shows a staccato of two units length. The unit here is the length of the typing moment of one character subject to the staccato effect. Staccato not only increases the specificity of the typing pattern, but also has the advantage of making the spacing between characters even. The typing time length of a character by a general typing method is 100ms to 130ms, whereas the typing time length of a character to which staccato is applied is 50 to 80ms.

6G is a typing pattern using legato. Legato is the opposite of staccato, where you type a letter as long as possible. The typing time length of the legato applied characters is generally about 350 to 400 ms. In general, the typing pattern to which the staccato and legato are applied has an advantage of higher consistency than the typing pattern by pose, but has a disadvantage of less change.

FIG. 6H illustrates a typing pattern using slow tempo. Slow tempo is to type the characters you type in the normal length, and to increase the spacing between the characters. That is, the user may increase the specificity of the typing pattern by increasing the typing interval. However, a typing pattern to which an excessively increased slow tempo is applied has a disadvantage of decreasing consistency.

In order to instruct the user to input the password in the above five ways, the user authentication device preferably displays a guide to the user for guiding such a typing method to the user.

Table 1 summarizes the characteristics of the typing pattern using the five artificial rhythms described above.

TABLE 1

advantage Disadvantages pose High specificity (diversity) Longer poses reduce consistency. Music rhythm Consistency is relatively high compared to others. You need a musical rhythm. Staccato Consistency is high. Specificity is limited. Legato Consistency is high. Specificity is limited. Slow tempo High specificity. Low consistency

That is, the typing pattern of FIGS. 6B-6H has an increased diversity compared to the typing pattern of FIG. 6A. Thus, the specificity of the user's typing pattern increases.

FIG. 7 is a diagram illustrating the quality of a pattern generated in the same manner as FIGS. 6B to 6H.

The measurement was performed for a six-digit password. In FIG. 7, inconsistency was used to measure consistency.

According to FIG. 7, it can be seen that the specificity increased from 200% (pose) to 500% (slow tempo) as in the case of long pause and slow tempo. In general, it can be seen that the discrimination is greatly increased compared to the natural typing method. In particular, it should be noted that the distinction of natural methods was negative. When the artificial rhythm is not applied to the discrimination (right y-axis) is a negative value, the discrimination after the artificial rhythm is applied all positively increased, the amount is also very large. The discrimination degree 0 represents a threshold corresponding to the design error rate, and the user authentication apparatus regards the user authentication as a success only when the difference between the circular pattern and the typing pattern is smaller than the threshold.

Thus, the typing pattern by artificial rhythm increases specificity and discrimination.

8A to 8H are diagrams illustrating a distance between the circular pattern, the test pattern, and the impersonator pattern in FIG. 7.

8A to 8H show the measurement results of FIG. 7 in more detail, a dotted line indicates a training pattern, a thin solid line indicates a test pattern, and a thick solid line indicates an impersonator pattern. By definition of each pattern, specificity, consistency and discrimination, the horizontal distance between the dashed line and the thick solid line intuitively represents the specificity (the greater the specificity, the larger the distance), and the horizontal distance between the dashed line and the thin solid line intuitively The greater the distance, the lower the consistency and increase the inconsistency.The distance between the thin solid line and the thick solid line intuitively represents the discrimination (the greater the distance, the greater the distinction).

8A to 8H, it can be seen that the thick solid line is shifted to the right side as compared to FIG. 8A to FIG. 8A. That is, the specificity increased. In addition, in the case of the long pose (FIG. 8C) and the slow tempo (FIG. 8H), it can be seen that not only the thick solid line is shifted to the right but also the thin solid line is shifted to the right. In other words, consistency decreased. In summary, it can be seen that typing patterns with artificial rhythms all increase specificity, but in some cases consistency tends to decrease. Timing queues in accordance with the present invention increase reduced consistency.

9A and 9B are diagrams comparing a conventional pattern with a pattern according to the present invention.

9A illustrates a pattern generated according to a conventional pattern generation method, that is, without rendering of a timing queue. As described with reference to FIG. 8H, the pattern generated using slow tampo increased the specificity of the pattern but decreased the consistency. In FIG. 9A, the horizontal distance P is an intuitive diagram of reduced consistency.

FIG. 9B is a pattern generated by a user sensing a rendering of a timing queue and inputting a password, and it can be seen that the horizontal distance P is almost zero compared to FIG. 9A. That is, the user can enter a password in a more consistent behavior pattern by keystrokes while watching or listening to the timing queue.

As shown in FIGS. 9A and 9B, it is common that both specificities were increased compared to FIG. 8A.

Applicants have obtained experimental data on the relationship between timing cues and specificity, consistency and discrimination to demonstrate the effectiveness of timing cues. In the experiment, five users entered the password “pass____word____” (using “pose” as an artificial rhythm) while listening to or watching the timing cues of the audio, video and audio / video signals. Table 2 shows the results of this experiment. .

TABLE 2

A metronome of 160 ticks / minute was used as the audio timing cue, and the video clip shown in FIG. 5 was used as the video timing cue. In the audio / video timing queue, these audio and video signals are synchronized and rendered to the user simultaneously.

For the first user, the inconsistency was reduced by 10 to 70 times without changing the specificity. As a result, the discrepancy increased significantly by about 7 to 12 times. Especially for this user, the audio / video timing cue proved to be the most effective.

For the second user, video timing cues reduced inconsistency, but did not increase discrimination significantly. This is because the second user has a very high degree of discrimination even in the absence of a timing queue. Especially in the case of audio timing cues, it turned out to be rather confusing.

For the third, fourth and fifth users, all three timing queues reduced the user's inconsistency. I think that these three users are the most common situation. It is determined that the type of timing queue positively affected by each user is different.

10 graphically illustrates the results of Table 2.

Figure 10 shows the specificity, inconsistency and discrimination for five users in the above experiment. The specificity by the video timing cue was measured up to 2086 but only up to 2000. As shown in FIG. 10, timing cues reduce overall inconsistency and increase discrimination of a user's pattern.

11 is a view showing a pattern generating apparatus according to an embodiment of the present invention.

The pattern generating apparatus 1100 includes a timing queue rendering unit 1110 and a timing vector calculator 1120.

The timing queue rendering unit 1110 renders to the user a timing queue 1112 which provides a regular signal to be a reference timing when a user inputs a password. The user strokes a password on the keyboard while sensing the timing cue 1112 visually or audibly. The pattern of the password that is input is a pattern that is increased because it is a pattern that is input after sensing a signal that is regularly provided.

The timing vector calculator 1120 generates a user pattern 1122 representing a user's typing pattern by calculating a timing vector of the input keystroke 1102 based on the timing queue.

12 illustrates a user authentication device according to an embodiment of the present invention.

The user authentication device 1200 includes a pattern generator 1210 and a determiner 1240.

The pattern generator 1210 generates the pattern 1222 by rendering the timing queue 1232 to the user, receiving the keystroke 1202 from the user, and calculating a timing vector. The pattern generator 1210 includes a timing queue renderer 1230 and a timing vector calculator 1220.

The timing queue rendering unit 1230 renders to the user a timing queue 1232 indicating regular timing. The timing cue can be an audio signal, a video signal, or a combination thereof, and in addition, includes all forms of representation that a human can feel.

In one embodiment of the invention, the timing queue rendered when the user enters the password is characterized in that the same timing pattern used when registering the user, i.e., generating the circular pattern. This allows the user to enter a password while listening to (or watching) a regular signal that was previously heard (or seen) at the time of registering the password, thereby further increasing the consistency of the pattern, thereby causing an error in the user authentication device. The rate can be further reduced.

The timing vector calculator 1220 receives the keystroke 1202 input by the user, and calculates a timing vector by measuring the elapsed time for each key input and the free time, thereby generating the pattern 1222.

The determination unit 1240 determines whether the user is a true user based on the user pattern 1222, the training pattern 1224, and the threshold value 1226. The determination unit 1240 includes a classification reference value calculation unit 1250 and a comparison unit 1260.

The determination unit 1240 is generally called a classifier, and generates a classification reference value by applying a predetermined function to a training pattern of a true user and a user pattern input when authenticating a user, and compares the classification reference value with a threshold value. Determine whether to authenticate.

The classification reference value calculator 1250 calculates the classification reference value 1252 using the user pattern 1222 and the training pattern 1224.

The classification reference value is a function value representing the relationship between the true user's pattern and the user's pattern input during authentication, and various values such as the difference between the circular pattern and the user pattern, the vector distance, or the Euclidean distance may be used.

The circular pattern is obtained from a training pattern which is a pattern of a password entered by the user at the time of password registration. For example, the circular pattern is an average value of the training patterns. The training pattern or the circular pattern may be stored in advance in a storage device (not shown) inside or outside the user authentication device 1200.

In one embodiment, when the circular pattern is (1,2,3,4) and the user pattern is (3,4,5,6), the classification reference value, i.e., the difference value of the vector is.

The comparison unit 1260 compares the classification reference value 1252 with the threshold value 1226, and if the classification reference value 1252 is larger than the predetermined threshold value 1226, authentication fails, otherwise authentication indicating success of authentication is performed. Generate a result 1262.

The threshold 1226 is a value determined by the user and has a very important influence on the accuracy or the error rate of the user authentication apparatus 1200. Therefore, determining the threshold is a very difficult task. However, since the pattern 1222 generated according to the pattern generation method according to the present invention is a pattern with increased consistency, the range of the threshold for achieving the desired error rate can be relaxed.

13 illustrates a user authentication method according to an embodiment of the present invention.

In operation 1310, the user authentication device renders a timing queue.

In step 1320, the user senses the timing queue and enters a password accordingly.

In operation 1330, the user authentication apparatus generates a pattern by receiving a keystroke according to a password and generating a timing vector.

In operation 1340, the user authentication apparatus calculates a difference between the pattern generated in operation 1330 and the circular pattern.

In step 1350, the user authentication device compares the difference value with the threshold value, and in steps 1360 and 1370, if the difference value is greater than the threshold value, the authentication is regarded as an authentication failure, otherwise, the authentication is completed and the procedure ends.

Meanwhile, the pattern generation method and the user authentication method according to the present invention can be created by a computer program. Codes and code segments constituting the program can be easily inferred by a computer programmer in the art. In addition, the program is stored in a computer readable media, and read and executed by a computer to implement a pattern generation method and a user authentication method. The information storage medium includes a magnetic recording medium, an optical recording medium, and a carrier wave medium.

So far I looked at the center of the preferred embodiment for the present invention. Those skilled in the art will appreciate that the present invention can be implemented in a modified form without departing from the essential features of the present invention. Therefore, the disclosed embodiments should be considered in descriptive sense only and not for purposes of limitation. The scope of the present invention is shown in the claims rather than the foregoing description, and all differences within the scope will be construed as being included in the present invention.

As described above, according to the present invention, a pattern with improved quality is generated by increasing the consistency. Further, according to the present invention, since the quality of the pattern and the circular pattern is increased, it is easy to determine the threshold of the user authentication device, and as a result, the error rate of the user authentication device is reduced.

Claims (12)

A method for generating a user pattern representing a password typing pattern of a user, used in a user behavior based user authentication system, (a) rendering a timing queue to the user that provides a regular signal to be a reference timing at the start of password entry of the user; (b) receiving a keystroke corresponding to the password input according to the timing queue; And (c) generating the user pattern by calculating a timing vector for the keystrokes. The method of claim 1, And the timing queue is one of an audio signal, a video signal, or a combination thereof. The method according to claim 1 or 2, And the user pattern is used in a classifier for the user authentication system. An apparatus for generating a user pattern representing a user's password typing pattern, which is used in a user behavior based user authentication system, (a) a timing queue rendering unit for rendering a timing queue for providing a regular signal to a user so as to become a reference timing when the user starts inputting a password; and (b) a timing vector calculator configured to generate the user pattern by calculating a timing vector for a keystroke of a user input according to the timing queue. The method of claim 4, wherein And the timing queue is any one of an audio signal, a video signal, and a combination thereof. The method according to claim 4 or 5, And the user pattern is used in a classifier for the user authentication system. In the user behavior based user authentication device, (a) a timing queue rendering unit for rendering a timing queue for providing a regular signal to a user so as to become a reference timing when the user starts inputting a password; a timing vector calculator configured to generate a user pattern representing a user's typing pattern by calculating a timing vector of an input keystroke based on the timing queue; And and (c) a determination unit that determines whether to authenticate the user based on the user pattern, a training pattern representing a true user pattern, and a threshold value. 8. The apparatus of claim 7, wherein the timing queue is the same as the timing queue used at the time of generating the training pattern. The method of claim 7, wherein And the timing queue is any one of an audio signal, a video signal, and a combination thereof. In the user behavior based user authentication method, (a) rendering a timing queue to the user that provides a regular signal to be a reference timing at the start of password entry of the user; (b) generating a user pattern representing a user's typing pattern by calculating a timing vector of an input keystroke based on the timing queue; And (c) determining whether the user is authenticated based on the user pattern, a training pattern representing a true user pattern, and a threshold value. 11. The method of claim 10, wherein the timing queue is the same as the timing queue used at the time of inputting the training pattern. The method of claim 10, And the timing queue is one of an audio signal, a video signal, or a combination thereof.

Images