With the development of computer network technology and e-commerce, the self-service banking system has got extensive popularization with the characteristic offering of high-quality 24 hours service for customer. Nowadays, using an ATM (Automatic Teller Machine) which provide customers with the convenient banknote trading is very common. However, the financial crime case has risen in recent years. A lot of criminals tamper with the ATM terminal and steal user's credit card and password by illegal means. Once user's bank card is lost and the password is stolen, the criminal can draw all cash in the shortest time, which will bring enormous financial losses to customer. How to carry on the valid identity becomes the focus in current financial circle. For the traditional ATM terminal customer recognition systems only rely on bank cards, passwords, and such identity verification methods are not perfect and functions are too single. Our new ATM terminal system is proposed and designed for solving the bugs of traditional ones. In this system additional to normal pin and password one more biometric technology, fingerprint is also added to provide high secured authentication. In recent years, the algorithm of fingerprint recognition is continuously updated, which has offered us a new verification means for us. The original password authentication method combined with the biometric identification technology verify the clients' identity better and improve the safety of ATM machines effectively. After entering the pin number, the user must give his finger print as another key process. If the fingerprint matches with the fingerprint which is already stored in the bank server's database, then only he can perform further transactions. The Token is generated by the Server to the user's mobile number. So that the legitimate user only can give the Token for the further process of authentication. Main advantage of our system is that it provides full security to the customer as well as it alerts the police if any suspicious person tries to use the ATM card.
THE CHARACTERISTICS OF THE HARDWARE DESIGN
The embedded ATM client authentication system is based on fingerprint recognition which is designed after analyzing the existed ATM system. The S3C2440 chip is used as the core of these embedded systems which is associated with the technologies of fingerprint recognition and current high speed network communication. The primary functions are shown as follows:
• Fingerprint recognition: The masters' fingerprint information was used as the standards of identification. It must certify the feature of the human fingerprint before using ATM system.
• Remote authentication: System can compare current client's fingerprint information with remote fingerprint data server.
• Telephone alarming: Once an exception happens, such as log in as the fake identity, the system will start the phone alarm to inform client and bank staff as soon as possible.
• Message alarming: the message can be send to the relevant staff's mobile phone without any noise, in order to carry on emergency processing.
• Police network connection: The system can call the police via the police network.
• Two discriminate analysis methods: Besides the fingerprint recognition, the mode of password recognition can be also used for the system.
The S3C2440 chip is used as the core of entire hardware.
Furthermore, the modules of LCD, keyboard, alarm, fingerprint recognition are connected with the main chip (S3C2440).The SRAM and FLASH are also embodied in the system. There are some modules consisted of the system as
LCD module: The OMAP5910 is used in this module as a LCD controller, it supported 1024*1024
image of 15 gray-scale or 3375 colors.
Keyboard module: It can be used for inputting passwords.
Alarming module: TC35i alarming module is based on GSM technology implement which can call the credit card owner and send message to relevant staffs without any sound.
SRAM and FLASH: The 16-bit 29LV160BB- 70REC of FLASH chip and the 32-bit HY57V561620CT-6 of SRAM chip are connected with the main chip. Their functions are storing the running code, the information of fingerprint and the algorithm.
Fingerprint recognition module: Atmel Company's AT77CI04B be used for fingerprint recognition. It has a 500dpi resolution, anti-press, anti-static, anticorrosion.
Ethernet switch controller: RTL8308B can provides eight 10/1 00 Mbps RMII Ethernet ports, which can connect police network and remote fingerprint data server.
Before using the ATM terminal, the client's fingerprint feature will be connected to the remote fingerprint data server to match fingerprint data with the master's, if the
result isn't correct, the system will call police automatically and send alarm to the credit card owner.
AT77C I 04B
S3C2440This is how the block diagram of the hardware module looks like.
S3C2440 works as a core of entire hardware.
AT77CI04B is used for fingerprint recognition and RTL8308B is used to connect to police network and server in which the fingerprints are saved.
The design of software is very important for this embedded system. The design included the design of main program flow chart, the
Initializing ones, and the algorithm of fingerprint recognition flow chart.
This system of software is implemented by the steps as follows: first of all, the Linux kernel and the File system are loaded into the main chip. The next, the system is initialized to implement specific task, such as checking A TM system, GSM communication and so on, and then each module reset for ready to run commands.
Before using A TM terminal, the password and fingerprint is required. First owner's password is entered and if password Matches with the password already saved
In the database, then the system requests for owner's fingerprint. If the fingerprint
matches with the fingerprint saved in the bank, the system would enter into the waiting status. In addition, the number of times the recognition of fingerprint and password are restricted to 3. If more than 3 times, the system will call the police through police network, telephone to the owner and
Send message to relevant staff. Then it locks the owner's credit card.
Another security measure which can be added is that if the password and fingerprint matches a token number can be generated and sent to the owner mobile. Now the owner must enter that token number and proceed with his transaction. This is an additional security measure which can be provided to the customer.
In the process of inputting fingerprint, the AT77CI04B which is a linear sensor captures fingerprint images by sweeping the finger over the sensing area. This will be used for acquiring the image of fingerprint. This product embed true hardware based 8-way navigation and click functions. The
fingerprint information will be temporarily stored in SRAM and upload to the remote fingerprint data server to compare
through bank network. The result of process will be controlled by main chip(S3C2440).
The initializing process means that set the hardware and software and then starts the multiple mission modules, each
Module will be started according to the priority processes. At first the system clock is initialized, and it executes the codes of
open interrupt and the open interrupt task. Then, the system would judge and enter process of module. Finally, the system would start to attempt multiple tasks.
THE DESIGN OF FINGERPRINT
The design of algorithm based on fingerprint recognition is so vital for the whole system. We would approach two steps to process the images of fingerprint.
Fingerprint recognition process
The first step is the acquisition of fingerprint image by the above device mentioned in the algorithm, and the results
could be sent to the following process.
Secondly, pre-processing the images acquired. After obtaining the fingerprint image, it must be pre-processed.
Generally, pre-processing is done by filtering, histogram computing, image enhancement and image linearization.
Lastly, the characteristic value was extracted, and the results of the above measures would be compared with the
information of owner's fingerprint in the database so as to verify whether the character is matched, and then the system
returns the results if it matches or not.
Fingerprint image enhancement
Fingerprint recognition module is an extremely important part of the system.
The high-quality images is the major
Factor for influencing the performance in the system. There is a lot of noise in fingerprint image. Now the image enhancement is the precondition for recognition of fingerprint characteristics. The algorithm of fingerprint recognition based on the algorithm of Gabor and direction filter is used. Fingerprint enhancement algorithm based on Gabor
Filter could be better to remove noise, strengthen the definition between the ridge and valley. It could significantly improve the image enhancement processing capacity, but this algorithm was slow in dealing with the high capacity requirements. Fingerprint enhancement algorithm based on direction filter has a faster processing capability, but it was not good in handing the large noise
Areas. So combination of these two algorithms could obtain better effects. The algorithm based on direction filter was
used in the clear area, and based on Gabor filter was used in the recoverable region.
The design of ATM terminal system based on fingerprint recognition has more advantages. It has advantages of the stability and reliability of fingerprint characteristics, a new biological technology. Fingerprints are unique for every individual and thus it provides high authentication. This is also based on the image enhancement algorithm of Gabor and direction filter. Additionally, this system also contains the original verifying methods which were inputting owner's password.
One drawback of this system is that in certain situation the owner is not able to go directly to the ATM and he sends some authorized user to withdraw money. It is not possible for that person to withdraw money because of fingerprint method. Here we have a solution for this situation. In addition to the owner's fingerprint another authorized user's fingerprint of the owner's choice can be added to the database. Now in a situation where the owner is not able to go to ATM, that user can use their fingerprint and withdraw money.
The security features were enhanced largely for the stability and reliability of owner recognition. The whole system was build on the technology of embedded system which makes the system more safe, reliable and easy to use.