Authentication of a user based on analyzing touch interactions with a device
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for continuous identity authentication based on gesture data. In one aspect, a method includes receiving gesture data corresponding to one or more gestures received on a touch screen of the mobile device from an individual, generating specific features extracted from the one more gestures, comparing the specific features with gesture data collected on the mobile device from a user previously interacting with the mobile device, and verifying that an identity of the individual matches an identity of the previous user.
1. A computer-implemented method for passive continuous identity authentication of an individual based on gesture data, the method being executed by one or more processors and comprising:
activating a continuous authentication mode to authenticate a user as they utilize swiping gestures on a touch screen;
training, by a mobile device, at least two clusters of classifiers for identifying a user by motion data and touch data features unique to the individual in a way the user executes a standard mobile device swipe gesture with each cluster of classifiers corresponding to a different orientation configuration of the user with respect to the mobile device selected from the group consisting of the user standing, walking, sitting, prone, inclined, and lying down;
receiving, by the mobile device, gesture data post sign-in for one or more swiping gestures, received on a touch screen of the mobile device from an individual during continued interaction by the individual with an application running on the mobile device;
generating, by the mobile device, specific features extracted from the one more swiping gestures including features of gesture patterns specific to the user for implementing a swipe gesture from the group consisting of a standard swipe right gesture, a standard swipe left gesture, a standard swipe up gesture, and a standard swipe down gesture;
comparing, by the mobile device, the specific features with the at least two clusters of classifiers stored on the mobile device; and
verifying, based on the comparing, that an identity of the individual matches an identity of the user.
2. A computer-implemented method, comprising:
activating a continuous authentication mode to authenticate a user as they utilize swiping gestures on a touch screen;
in an enrollment phase, training, by a mobile device, a first set of classifiers for identifying a user by features unique to touch data and motion data in a way the user executes a standard mobile device gesture in a first orientation configuration of the user with respect to the mobile device, including training a first classifier for a standard swipe right gesture of the user, a second classifier for a standard swipe left gesture of the user, a third classifier for a standard swipe up gesture of the user, and a fourth classifier for a standard swipe down gesture of the user;
updating the training to train a second set of classifiers for identifying the user by features unique to the touch data and motion data in the way the user executes a standard mobile device gesture in a second orientation configuration of the user with respect to the mobile device, including training a fifth classifier for a standard swipe right gesture of the user, a sixth classifier for a standard swipe left gesture of the user, a seventh classifier for a standard swipe up gesture of the user, and an eighth classifier for a standard swipe down gesture of the user;
receiving gesture data including motion data and touch data post sign-in for swiping gestures received on a touch screen of the mobile device from an individual during continued interaction by the individual with an application running on the mobile device;
generating specific motion and touch features extracted from including motion and touch features of gesture patterns specific to the individual for swiping gestures from the group consisting of the standard swipe right gesture, the standard swipe left gesture, the standards swipe up gesture, and the standard swipe down gesture;
comparing, by the mobile device, the specific features with the first and second set of classifiers stored on the mobile device;
verifying that an identity of the individual matches an identity of the user wherein the touch data and the motion data includes location, force, and size of the one or more swiping gestures; and
wherein the first and second orientation configurations are different orientation configurations selected from the group consisting of a user standing, walking, sitting, prone, inclined, and lying down.