Method of using SVM classifier to identify handset users

Abstract

The invention discloses a method of using an SVM classifier to identify handset users. Firstly, a user holds a handset to make a prescribed motion and through a handset linear acceleration sensor and a direction sensor, the values of a three-dimension coordinate (x, y, z) are acquired for once in a sampling period and the obtained data is processed and then the SVM is used to carry out classification identification. After the processing is completed, different users can be distinguished accurately as long as the users make specified motions so that a user identification objective is achieved. The method is convenient and rapid and high in security.

Description

technical field [0001] The invention relates to identification technology, in particular to a method for extracting physiological features using mobile phone sensors and classifying users with SVM (Support Vector Machines). Background technique [0002] The SVM (Support Vector Machines, Support Vector Machines) classifier in mobile phone user identification is a general machine learning method based on the statistical learning theory framework. It was originally proposed for two types of classification problems. It has a simple structure and generalization ability. Strong advantage. [0003] With the rapid development of e-commerce, more and more people expect to realize mobile e-commerce such as mobile payment, mobile coupons, mobile banking and mobile VIP customer service terminal, as well as mobile transaction services based on mobile terminals, through the mobile phone platform. This makes mobile phones face serious security challenges. [0004] At present, the mobile ...

AI Technical Summary

Problems solved by technology

This greatly weakens the user experience level

Moreover, password-based authentication schemes are vulnerable to external malicious attacks, and there are security risks such as leakage, forgery, and forgetting of user names or passwords.

[0005] The increasingly mature physiological feature biometrics is based on the measurement data of human characteristics, which are direct anthropometric features, such as face, fingerprint, iris, etc., but face, fingerprint, iris and other recognition technologies are image-based recognition technologies, there The risk of being counterfeited can only be used to identify the real identity of the user, but cannot prove that the verification image is the real-time collected image of the mobile phone user himself.

Images