Method for generating a human likeness score

• US 9,501,630 B2
• Filed: 05/30/2014
• Issued: 11/22/2016
• Est. Priority Date: 03/24/2011
• Status: Active Grant

• Current Assignee: Imperva Incorporated (Thales SA)
• Original Assignee: Ayah LLC (Thales SA)
• Inventors: Paxton, Tyler James, Tatoris, Reid Michael, Blackmer, Benjamin Crispell, Aleska, Adam, VandenBrink, Stuart
• Primary Examiner(s): SHAIFER HARRIMAN, DANT B
• Application Number:
  US14/292,266
• Publication Number:
  US 20150347731A1
• Time in Patent Office: 907 Days
• Field of Search: 726/2
• US Class Current: 1/1
• CPC Class Codes: G06F 16/24578 using ranking
• CPC Class Codes: G06F 21/31 User authentication
• CPC Class Codes: G06F 21/316 by observing the pattern of computer usage, e.g. typical user behaviour
• CPC Class Codes: G06F 21/32 using biometric data, e.g. fingerprints, iris scans or voiceprints
• CPC Class Codes: G06F 21/36 by graphic or iconic representation
• CPC Class Codes: G06F 2221/2103 Challenge-response
• CPC Class Codes: G06F 2221/2133 Verifying human interaction, e.g., Captcha
• CPC Class Codes: H04L 63/083 using passwords cryptographic mechanisms or cryptographic arrangements for entity authentication using a predetermined code H04L9/3226

Abstract

One embodiment of the invention is a method utilizing a CAPTCHA to generate a human likeness score including blocks: a) receiving a user solution to the CAPTCHA; b) receiving a user interaction pattern descriptive of an interaction undertaken by the user, through a graphical interface of the CAPTCHA, to achieve the user solution; c) determining the accuracy of the user solution; d) comparing the user interaction pattern against an interaction model generated from interaction patterns of previous users; e) calculating the human likeness score based upon the determination of block c) and the comparison of block d), wherein the human likeness score lies within a continuum of human likeness scores.

Claims

1. A method for generating a human likeness score based on a Completely Automated Public Turing test to tell Computers and Humans Apart (CAPTCHA), the method comprising:

selecting a visual advertising asset;

generating the CAPTCHA comprising a graphical interface, the visual advertising asset and a graphical object depicted within the graphical interface;

assigning an instruction for completing the CAPTCHA;

from a user, receiving an input within the graphical interface, the input comprising a solution to the CAPTCHA from the user, the solution comprising a selection of the graphical object within the graphical interface;

extracting a motion pattern from the input;

comparing the motion pattern to a motion model based on input patterns of previous users; and

calculating the human likeness score of the user based on the comparison of the motion pattern to the motion model, the human likeness score lying within a continuum of discrete human likeness scores, wherein calculating the human likeness score comprises;

generating a first confidence group according to an accuracy of the solution in fulfilling the instruction;

generating a second confidence group based on the comparison of the motion pattern to the motion model;

generating a third confidence group based on user data pertaining to at least one of a CAPTCHA previously attempted by the user, an Internet Protocol address of the user, and a cookie associated with the user;

determining a reliability of the first confidence group, the second confidence group, and the third confidence group;

compiling the first confidence group, the second confidence group, and the third confidence group, based on the determined reliability thereof, to generate the human likeness score; and

selecting a number value, from the continuum of human likeness scores comprising a continuum of number values, that correlates with a calculated confidence that the user is human.

2. The method of claim 1, wherein generating the CAPTCHA comprises generating the CAPTCHA including the visual advertising asset and the graphical object that are visually coextensive within the graphical interface.

3. The method of claim 1, wherein generating the CAPTCHA comprises inserting a two-dimensional representation of an advertised product within the graphic interface.

4. The method of claim 1, wherein selecting the visual advertising asset comprises retrieving the visual advertising asset from a publisher, and further comprising passing the human likeness score of the user to the publisher.

5. The method of claim 1, wherein receiving the input within the graphical interface comprises recording a path of a cursor within the graphical interface, wherein extracting the motion pattern from the input comprises characterizing a geometry of the path of the cursor, wherein comparing the motion pattern to the motion model comprises comparing a geometry characteristic of the path of the cursor to an input geometry model generated from an input previously submitted by a known human, and wherein calculating the human likeness score of the user comprises associating the user with a human likeness score based on a congruence between the geometry characteristic of the path of the cursor and the input geometry model.

6. The method of claim 1, wherein extracting the motion pattern from the input comprises isolating a noise component from the input, wherein comparing the motion pattern to the motion model comprises comparing the noise component of the path of the cursor to an input geometry model generated from an input previously submitted by a known human, and wherein calculating the human likeness score of the user comprises calculating a divergence of the noise component from a human input noise model.

7. The method of claim 6, wherein extracting the motion pattern from the input comprises extracting the noise component from a touch input on a portion of a touchscreen corresponding to the graphical interface.

8. The method of claim 1, wherein selecting the visual advertising asset comprises selecting the visual advertising asset to distinguish the CAPTCHA from a second CAPTCHA previously served to the user.

9. The method of claim 1, wherein generating the CAPTCHA comprises assigning a motion characteristic to the graphical object to differentiate the CAPTCHA from a previously-generated CAPTCHA and initiating motion of the graphical object within the graphical interface based on the motion characteristic.

10. The method of claim 9, wherein receiving the input within the graphical interface comprises recording a path of a cursor within the graphical interface, wherein extracting the motion pattern from the input comprises characterizing a lag in the path of the cursor relative to motion of the graphical object within the graphical interface, and wherein calculating the human likeness score of the user comprises quantifying a similarity between the lag in the path of the cursor and an input lag model.

11. The method of claim 1, wherein generating the CAPTCHA comprises setting a number of graphical objects for the CAPTCHA and depicting the number of graphical objects comprising the graphical object within the graphical interface.

12. The method of claim 1, wherein assigning the instruction for completing the CAPTCHA comprises assigning, to the CAPTCHA, an instruction related to at least one of a shape, the color, the size, the type, and the context of the first object within the graphical interface.

13. The method of claim 1, wherein receiving the input comprises receiving a keystroke entered by the user, and wherein comparing the motion pattern to the motion model comprises comparing a key state and a timing of the keystroke to the motion model.

14. The method of claim 1, further comprising predicting that the user is human based on the human likeness score that is greater than a threshold likeness score, predicting that the user is automated software based on the human likeness score that is less than the threshold likeness score, and adding the motion pattern to a database of motion patterns in response to predicting that the user is human, the database comprising input patterns of previous users verified as human.

15. The method of claim 1, wherein calculating the human likeness score of the user comprises distinguishing the user from a second user based on the human likeness score.

16. A method for generating a human likeness score based on a Completely Automated Public Turing test to tell Computers and Humans Apart (CAPTCHA), the method comprising:

based on a target difficulty of the CAPTCHA, setting a number of objects for the CAPTCHA;

assigning an instruction for completing the CAPTCHA;

generating the CAPTCHA comprising a graphical interface and a set of graphical objects depicted within the graphical interface, the set of graphical objects corresponding to the number of objects for the CAPTCHA;

from a user, receiving an input within the graphical interface responsive to a graphical object in the set of graphical objects within the graphical interface, wherein receiving the input comprises receiving a solution to the CAPTCHA from the user;

extracting an input pattern from the input; and

calculating the human likeness score of the user based on a comparison of the input pattern to an input model, the input model based on an input pattern of a previous user, and the human likeness score lying within a continuum of discrete human likeness scores, wherein calculating the human likeness score comprises;

generating a first confidence group according to an accuracy of the solution in fulfilling the instruction;

generating a second confidence group based on the comparison of the input pattern to the input model;

generating a third confidence group based on user data pertaining to at least one of a CAPTCHA previously attempted by the user, an Internet Protocol address of the user, and a cookie associated with the user;

determining a reliability of the first confidence group, the second confidence group, and the third confidence group;

compiling the first confidence group, the second confidence group, and the third confidence group, based on the determined reliability thereof; and

selecting a number value, from the continuum of human likeness scores comprising a continuum of number values, that correlates with a calculated confidence that the user is human.

17. The method of claim 16, wherein generating the CAPTCHA comprises initiating motion of a particular graphical object in the set of graphical objects within the graphical interface, and wherein receiving the input within the graphical interface comprises recording a cursor input into the graphical interface to select the particular graphical object within the graphical interface.

18. The method of claim 17, wherein initiating motion of the particular graphical object pseudorandomly assigning at least one of an initial location, an initial orientation, an initial speed, and an initial direction of motion to the particular graphical object within the graphical interface.

19. The method of claim 17, wherein extracting the input pattern from the input comprises characterizing a geometry of the cursor input, wherein calculating the human likeness score of the user comprises characterizing a geometry of the cursor input, comparing a geometry characteristic of the cursor input to an input geometry model generated from an input previously submitted by a known human, and associating the user with a human likeness score based on a congruence between the geometry characteristic of the cursor input and the input geometry model.

20. The method of claim 16, wherein setting a number of objects for the CAPTCHA comprises setting the number of objects for CAPTCHA to differentiate the CAPTCHA from a previously-generated CAPTCHA.