An identity verification system based on writing pen strength recognition

In today's information society, personal privacy, commercial secrets and national security are increasingly valued by people. Information needs to be both conveniently readable and reliably protected. However, some of today's identity authentication schemes are difficult to be perfect. Even iris and fingerprint recognition, which are considered to be unreplicable, can be impersonated in extreme cases when criminals obtain the eyeballs or fingers of the parties involved. Since the writing habits of adults, including the shape, speed, strength and angle of writing, are fixed for a long time, the shape of the signature has long been used as the basis for identity authentication. In particular, Chinese characters are often used for identity authentication because of their obvious characteristics. However, the shape of the characters is an intuitive form and is easier to imitate. In contrast, the strength of writing is difficult to detect and is almost impossible to copy and imitate. Due to its own structure, the resistive touch screen can use a dedicated touch screen control chip to collect handwriting and pen strength information. Therefore, this paper proposes a new way to authenticate identity-a verification system based on writing pen strength recognition.

In addition, the pen strength can be converted into the third dimension information and processed using the now quite mature image recognition algorithm. This system combines handwriting recognition technology with the authentication of writing strength to improve the accuracy of identity authentication. This system uses an embedded platform to basically realize the identity authentication based on writing pen strength, which is easy to transplant. With the increasing popularity of mobile devices and the increasing concern for personal privacy, this solution is likely to be widely adopted.

1 System overall design

This system completes the task of collecting the handwriting and pen strength information of the user's signature through the signal acquisition module, and then uses LabVIEW to implement software development, establish a pen strength information database, and finally complete the user's authentication through a matching and comparison algorithm.

The goal and function of this system are: to establish a signature system with high security, difficult to impersonate and copy, and widely used in various identity authentication. It adopts modular design as a whole as shown in Figure 1. The AD7879 chip, single-chip microcomputer and touch screen are used to complete the collection of pen strength data information, and then the data communication with the computer is completed through the serial port, and the pen strength information data is handed over to the host computer; the host computer uses LabVIEW to build a friendly user interface, and then the program body converts the obtained data, and after encrypted storage, it is displayed on the screen according to certain rules. At the same time, the background will complete the extraction of feature values, match them with existing data, and display the results to the user in the form of a dialog box.

Figure 1 System framework diagram

2 System Hardware Design

2.1 Pressure measurement principle

By analyzing the structure of the resistive touch screen, we can know that when the touch screen is pressed, there will definitely be contact resistance between the two layers of metal film, as shown in Figure 2. The larger the touch resistance value, the smaller the touch pressure, so here we can get a solution to measure touch pressure.

Figure 2 Pressure measurement principle diagram

The specific measurement method is as follows: connect the reference voltage to Y + and X -, measure the voltage at the X + terminal to get Z1 and the voltage at the Y - terminal to get Z2. The calculation formula for contact resistance is:

In the formula, RTOUCH represents the contact resistance, RX represents the total resistance in the X direction, which is related to the actual touch pressure, XPOS represents the voltage sampling value in the X direction at the contact point, and N represents the sampling depth. In this project, we only care about the relative size of the pressure, so we don’t need to know the specific value of RX, which ensures the portability of the software.

2.2 Working Principle of Touch Screen Controller AD7879 Chip

The AD7879 chip is a dedicated four-wire resistive touch screen driver chip that supports a single power supply of 1.6 ~ 3.6 V. Its main modules include voltage switching and channel selection circuits, internal 2.5 V reference voltage, 12-bit successive approximation analog-to-digital converter, data preprocessing module, serial digital interface and digital control logic. The voltage switching and channel selection circuits of the AD7879 chip can easily switch the reference voltage access terminal and ADC sampling terminal on the touch screen to obtain the required X, Y and Z1, Z2 values. In order to reduce the impact of LCD noise and other noise sources, the AD7879 also has a built-in preprocessing module, and the preprocessing functions include median filter and mean filter. The interface between the AD7879 and the microcontroller is a standard SPI serial interface or I2C interface (depending on the specific model). The chip also has an interrupt enable port. When there is pressure on the touch screen, the interrupt enable terminal immediately generates an interrupt signal to the MCU to start collecting coordinates.

2.3 AD7879 and MCU interface

The external interface circuit of the AD7879 chip is quite simple. It can work by simply connecting to a touch screen and a power supply. Then connect the serial interface (SPI or I2C) of the AD7879 to the MCU, as shown in Figure 3, to obtain the position coordinates and pressure information.

Figure 3 AD7879 typical interface circuit

3 System Software Design

3.1 Lower computer software design

The lower computer is a 80C51 single-chip microcomputer, which mainly controls the AD7879 chip to collect the font and pressure information on the touch screen, and sends the coordinate and pressure information to the upper computer program. As shown in Figure 4, the lower computer program flow chart, first the single-chip microcomputer initializes the AD7879 chip, and then waits for the external interrupt trigger. After the interrupt is triggered, the interrupt timer is turned on, and the start code is sent to inform the upper computer program to start writing, and then enters the loop to wait for the timer interrupt. After entering the timer interrupt, the writing position and pressure on the touch screen are immediately read and sent to the upper computer program. When there is no pressure on the touch screen, the writing is considered to be completed, and then the timer is turned off, the end character is sent and the external interrupt is exited. The use of timer interrupts ensures that the time interval between each two samples is the same.

Figure 4 Flowchart of the lower computer program

3.2 Host computer program

The host computer program is written using LabVIEW.

LabVIEW is a graphical programming platform launched by National Instruments. It introduces the concept of virtual instruments and writes codes in the form of block diagrams. Its high degree of modularity and intelligence facilitates software design.

To simplify development, this system uses LabVIEW for interface development and algorithm processing.

The identity authentication function of this system mainly includes two steps. The first step is to create a signature, that is, to record the writing characteristics of the user who is allowed to enter the system into the system; the second step is to verify the signature, that is, to verify whether the user who enters the signature has the authority to enter the system. The system block diagram is shown in Figure 5.

Figure 5 Flowchart of host computer program

The system is first initialized, the user selects "Create Signature", and then starts writing the signature on the touch screen. The system receives the data transmitted by the lower computer and stores it first. The received data is valid from the start code to the end code. Because the AD7879 has a 12-bit ADC inside, and each sampling point needs to collect data four times (X, Y, Z1, Z2), each sampling point corresponds to 8 Bytes of information. The upper computer program can read out all valid points in sequence, read each sampling point in a loop, and draw them point by point on the blank picture, and you can get the written font display picture. In order to show the display effect vividly, the pressure characteristics are mapped into the color information of the pixel points in the design. Dark colors indicate high pressure, and light colors indicate low pressure.

After the signature is created, it can be verified. The user selects the verification signature and starts writing on the touch screen. The data collection and processing process is the same as the method of creating the signature, and image B is obtained. Then, the input signature is compared with the initially created signature. The comparison process is as follows.

(1) Eliminate error points. Experiments have found that some sampling points in the data collected by the touch screen are caused by system random errors, which appear as some scattered isolated points on the screen. These error points need to be eliminated before verification.

(2) Align the characters. First, match the handwriting to each other. This is mainly to find the characteristic points in the handwriting, including the lifting point, the landing point and the turning point, and form characteristic marks.

(3) Segment characters. Based on the marks made in the previous step, the handwriting is divided into several segments with feature points as intervals, and compared with the original signature data created. If the number of segments is different, the verification fails and a re-verification prompt is sent. Otherwise, proceed to the next step of comparison.

(4) Pressure comparison. Calculate the difference in average pressure near each feature point. The sum of the curvature at each turning point and the mean square error of each sampling point sequence. If these results do not exceed the set threshold, the authentication is passed, otherwise the authentication fails. These thresholds are obtained through repeated comparisons in experiments. This threshold can also be set according to user needs. The higher the security level required by the user, the higher the threshold is set.

4 Results and discussion

By collecting writing pressure information through the hardware platform of this system, it can be analyzed that there are obvious differences in the writing pressure of different people.

Figure 6 shows the change in pen strength when writing the character "梅". Because the pressure becomes very small each time the pen is lifted, 11 peaks can be seen in the waveform, corresponding to the 11 times the pen is lifted. The writing strokes can be segmented based on the peaks. The speed of each writing is not perfect, so the number of samples for each stroke is also different. The interpolation algorithm is used to normalize the number of samples for each stroke to the same, so that the one-to-one correspondence effect of the sampling points can be seen in the image.

Figure 6 Pen force change curve

The gray line in the figure represents the pressure change collected when the same person writes, and the dark gray line and black line represent the pressure change when two other people write. The pressure change trend and absolute size of the pressure when the same person writes are very close, and can be distinguished from the writing of different people, indicating that the design idea of ​​this system is reasonable.

5 Conclusion

This paper introduces an identity authentication system based on pen strength recognition, focusing on the improvement of the previous method of identifying identity by simply using the shape of the written characters. Based on the relatively mature handwriting recognition technology, supplemented by the comparison of the writing pen strength, the user's identity can be more accurately identified. Moreover, the writing pen strength is difficult to detect in appearance, so it is difficult to imitate. This system uses LabVIEW to write the host computer, which fully simplifies the development process. In addition, the collection of writing pen strength in this article is directly based on the touch screen, without any external sensors and special pressure collection equipment. It is simple and easy. With a little modification, this solution can be widely used on other touch-screen-based devices. After actual testing, this system has a high recognition rate for specific users, but there is a certain rejection rate of authentication. This requires further improvement of the algorithm to further improve the accuracy. In practical applications, this method should be combined with handwriting recognition to improve the accuracy of authentication.