Smart device identity authentication method and client based on touch screen operation

Abstract

The invention provides a smart device identity authentication method based on touch screen operation and a client terminal, which improves the security of password input authentication. The present invention collects and records the user's behavior data during multiple touch screen operations; extracts the behavioral characteristics of the behavior data during multiple touch screen operations, and establishes behavioral feature vectors for multiple touch screen operations; The behavior feature vector of the operation is normalized and a classifier is constructed; the classifier is stored and sent to the cloud for storage; the behavior data of the user during a single touch screen operation is collected and recorded; the single touch screen operation process is extracted Behavior characteristics of the behavior data in the data, and establish the behavior feature vector of a single touch screen operation; normalize the behavior feature vector of the single touch screen operation; The behavior feature vector of touch screen operation realizes the implicit authentication of password input.

Description

technical field [0001] The invention belongs to the field of information security, and in particular relates to a touch screen operation-based smart device identity authentication method and a client. Background technique [0002] Smart devices have become ubiquitous personal devices. The usage coverage rate of smart devices in the United States has reached 89%, and sales continue to rise. Smart devices are convenient for users to chat online, transfer files, trade and shop anytime, anywhere. A large amount of sensitive private information is stored in smart devices, which brings great security risks to users, because smart devices contain many important private information, such as confidential Files, logins, contacts, etc. In addition, as a portable device, a smart device is more likely to be lost or stolen than a typical desktop electronic device. Therefore, how to ensure that private information is not stolen by illegal users is the focus of research by smart device d...

AI Technical Summary

Problems solved by technology

Although smart devices on the market are basically equipped with a digital password lock function to facilitate user authentication, there are several problems in the following aspects: First, many people choose passwords with low password strength in order to avoid forgetting passwords. According to the survey of four-digit code locks for devices, "0000", "1234" and "2580" have become the top three most popular passwords, accounting for 15% of the total number of people surveyed; secondly, even if a password with a low repetition rate is used, However, the password is easy to reveal. For example, in a crowded bus, if no security measures are taken when unlocking the lock, it is easy to be peeped by people around; Password, these problems make this kind of password lock useless, illegal users can successfully enter the user interface by entering the correct password, and then steal the privacy and confidential files in the user's smart device (such as browsing chat records, copying confidential information, stealing bank accounts) password, etc.)

Although the gyroscope is used to simply detect the distance feature of the user during the touch screen operation and provide an implicit authentication method for touch screen operation, but first of all, the distance feature is only a part of the touch screen operation process, and the simple distance feature is difficult for the user. The touch screen operation is completely detected; secondly, after extracting the behavior feature vector from the gyroscope behavior data, a single classifier is directly used to establish the user’s identity authentication model, and the extracted behavior feature vector is not processed by data standardization. The measured data does not all change according to certain rules. There will always be outliers in the data detected by the gyroscope during the user's use. If these outliers are not processed, the training model will fail to be established or the identity of the distance feature vector High authentication failure rate, which interferes with user touch screen operations

Images