Smart phone auxiliary authentication method based on multi-mode fusion

Abstract

The invention discloses a smart phone auxiliary authentication method based on multi-mode fusion. The existing continuous authentication method does not fully utilize abundant data generated when a user operates a mobile phone, the recognition accuracy is low, and the recognition performance is easily influenced by the body posture change of the user. According to the method disclosed by the invention, the identity of the user is continuously authenticated by utilizing the single-finger sliding operation, the continuous keystroke behavior and the sensor data within the duration of the two behaviors of the user. According to the method, a continuous keystroke behavior is set, a two-stage classification technology is adopted, user operation postures are divided into three postures, namely a hand-held mobile phone static posture, a hand-held mobile phone walking posture and a posture that the mobile phone is placed in a static object, the current posture of a user is recognized firstly, and then feature vectors are sent into classifiers under the corresponding postures. According to the method, samples from legal registered users are used for training six single classifiers, and the single classifiers are used for identity recognition. The method is rich in data source and high in robustness, and the authentication accuracy is improved.

Description

technical field [0001] The invention belongs to the technical field of information security, and in particular relates to a smart phone auxiliary authentication method based on multimodal fusion. Background technique [0002] With the widespread use of smart phones and the continuous enrichment of mobile phone functions, a large amount of personal privacy information such as social networking, pictures, voice and financial information is stored in mobile phones. If these private information is stolen, it may cause damage to users' property and even personal safety. major threat. Identity authentication of current users using mobile phones is an important means to protect personal privacy in mobile phones. Traditional authentication methods such as passwords, patterns, and emerging technologies such as fingerprint and face recognition have certain limitations, such as passwords are easy to forget and be stolen. Shoulder spying, the pattern is more likely to be snooped by a t...

AI Technical Summary

Problems solved by technology

This method uses a two-class support vector machine as a classifier, which is inappropriate, because in practice, the data of illegal users cannot be obtained in advance for training the model

In addition, this method does not consider the user's keystroke behavior for identity authentication. Keystroke behavior is one of the most common behaviors in the process of using a smartphone and contains a wealth of user personal information.

Images