In the future, banks and Government agencies could have the option of imposing a higher level of security clearance as a measure of curbing illegal access by unauthorised individuals.
We are not talking about biometric solutions like fingerprint-, voice- or face-recognition. While these solutions are deployed commercially, they are subjected to physical damage such as dry skin, loss of voice, scars, etc.
Instead we are looking at verifying a person's identity based on his or her brain waves or heartbeat signals. This is the research project that four students of Multimedia University (MMU) have been working on since Aug 2009 and which is expected to be completed this May.
The four students are Muhammad Kamil Abdullah, 25, Syed Syahril Syed Ibrahim, 26, Justin Leo Cheang Loong, 24 and Lam Zheng Yan, 23. The project is supervised by Dr Khazaimatol Shima Subari, who is a lecturer and researcher at the engineering faculty in MMU.
The students have developed a multi-modal biometric system called Bio-Secure that uses a person's brain waves (EEG) and heartbeat signals (ECG) to verify or authenticate an individual. Like the name suggests, a multi-modal biometric system uses more than one biometric technology to identify and verify a person.
In layman's terms EEG(electroencephalogram) is a measurement of the electrical activity of the brain produced by the firing of neurons (within the brain).
ECG (electrocardiogram) is used to measure the rate and regularity of a person's heartbeat.
"We believe Bio-Secure is Malaysia' first EEG biometric system. As far as we know there are no biometrics system based on EEG and ECG signals being developed in Malaysia," said Muhammad Kamil, who is pursuing his Masters in engineering science at the university.
Syed Syahril, who is also pursuing the same Masters degree, said EEG and ECG signals are unique among individuals and are hard to forge. Therefore, they can be used as an ideal form of biometric security.
About Bio-Secure
The Bio-Secure solution is an identification and verification biometrics system. It offers two options: offline and online processing.
"You choose the offline option when you want to identify or verify a person using a pre-recorded signal," said Leo.
Verification means confirming or denying the identity claimed by a person. In this mode, the system will state how well the pre-recorded signal matches with the one in the database.
An application that uses the verification mode, for example, is when you want to log into a computer, smartphone or computer network. In this case, you know the username and password to log into the computer.
"Another example is the verification process at the Kuala Lumpur International Airport. The automated passport lane uses our fingerprint to verify our identity. In Bio-Secure, the verification process is done using the EEG and ECG signals," said Khazaimatol, who has been lecturing on signal processing at MMU for four years.
The online processing option is used when you want to identify a person using EEG and ECG signals recorded in real time.
In the online processing option, Bio-Secure records the EEG and ECG signals of a person for a few seconds and then immediately compares this signal with all the pre-recorded signals (of all persons) stored in the database.
The system will select the person in the database who best matches the signal.
This option is suitable for applications such as information retrieval from a police database or video surveillance in public places.
"The accuracy rates for the multi-model EEG and ECG system is approximately at 88% to 90%," said Khazaimatol.
Additionally, Bio-Secure also offers another option called Training.
"We use this option when we need to add a new person to the database for identification and authentication purpose," said Leo, who is also studying for his Masters in engineering science at the university.
"Currently, the database consists of 35 pre-recorded EEG and ECG signals, which are from students of MMU," said Nicolas Lim, who is a electronics engineering undergraduate at the university.
He helps to test the Bio-Secure biometrics system.
WIRED UP: Electrodes are placed on Syed Syahril's scalp and wrists before taking the measurement of his brain waves and heartbeat.
Getting the EEG and ECG signals
At a first glance, the Bio-Secure may seem like a straightforward system but it takes at least 20 minutes to prepare a person to record his or her EEG and ECG signal.
To measure a person's EEG signal, four gold-plated electrodes are placed on several locations on the scalp.
The electrodes are smeared with gel to ensure that the metal bands at the bottom of the electrodes come into contact with the scalp.
As for measuring ECG signals, one electrode is placed on the person's neck and another two are placed on the underside of the right and left wrists respectively.
To record the EEG and ECG signals, the individual electrodes are connected to a recording console. Here, the signals are transmitted from the console to the computer, where they are processed using the Matlab package for scientific and engineering computation.
During the entire process, the person also needs to be calm and relaxed.
Not yet perfect
The Bio-Secure solution in its current state is still not a practical system as the recording of a person's EEG and ECG signals still takes a long time.
"This is because preparing the electrodes and attaching them on a person's scalp, neck and wrists takes time. If we can shorten these processes that will be great," said Lam, who added that scientists are still working on how to overcome these challenges.
"Previously we experimented with several locations (using 64 electrodes) on a person's scalp. After months of experiments, we managed to locate four positions that gave the most accurate readings," said Muhammad Kamil.
Khazaimatol said EEG and ECG signals are unique but there are only so many features one can use in the signal. Therefore, the success rate depends highly on the algorithms as well as the quality of the signal.
"It is difficult to get a 100% accuracy rate unless the sample size is between 10 to 20 people. Therefore, we have to test the Bio-Secure system on a larger database," she said.
"A larger database refers to collecting EEG and ECG signals from different data subjects (different races, groups, etc) using different stimulus. At the same time, we will improve the algorithm to ensure higher accuracy rates," she added.
The next step
When the project first started in 2009, Bio-Secure was funded by the Science, Technology and Innovation Ministry (Mosti).
The money went to the purchase of data acquisition devices, computers, and the hiring of research officers (ROs).
The ROs are the four students. Muhammad Kamil and Syed Syahril have worked on audio signal processing projects while Leo did a research on heart sound classification while Lam conducted research on ECG classification.
"The ROs involved in the Bio-Secure project can use their research findings in their Masters degree," explained Khazaimatol.
"Although Mosti has no more funding allocated for the Bio-Secure project this year, the ROs are still working on the project on part-time basis except for Syed Syahril, who is working on the project full-time," she added.
The students now want to develop their own biometrics hardware i.e. EEG and ECG data acquisition devices. This is because the current data acquisition devices used by the Bio-Secure system is from a company that sells medical equipment, and are not tailored according to their needs.
"This will be an entirely new project which will require further funding," said Khazaimatol.
More time needed
The Bio-Secure project is a good start towards developing our own biometrics system based on EEG and ECG signals.
However in its current state, the Bio-Secure is still not ready for commercialisation. More research, commitment and funding is still needed and the system may conceivably be a viable solution in the future.
If it takes off, the next time you walk in a bank you may very well be using a home-grown Bio-Secure biometrics security system.
No comments:
Post a Comment