Khairul Azami Sidek

Data mining in mobile ECG based biometric identification

This paper investigates the robustness of performing biometric identification in a mobile environment using electrocardiogram (ECG) signals. We implemented our proposed biometric sample extraction technique to test the usability across classifiers. Subjects in MIT-BIH Normal Sinus Rhythm Database (NSRDB) were used to validate the reliability and stability of the subject recognition methods. Discriminatory features extracted from the experimentations were later applied to different classifiers for performance measures based on the complexity of our proposed sample extraction method when compared to other related algorithms, the total execution time (TET) applied on different classifiers in various mobile devices and the classification accuracies when applied to various classification techniques. Experimentation results showed that our method simplifies biometric identification process by obtaining reduced computational complexity when compared to other related algorithms. This is evident when TET values were significantly low on mobile devices as compared to a non-mobile device while maintaining high accuracy rates ranging from 98.30% to 99.07% in different classifiers. Therefore, these outcomes support the usability of ECG based biometric identification in a mobile environment.

ECG Biometric with Abnormal Cardiac Conditions in Remote Monitoring System

This paper presents a person identification mechanism using electrocardiogram (ECG) signals with abnormal cardiac conditions in network environments. A total of 164 subjects were used in this paper using three different databases containing various irregular heart states from MIT-BIH arrhythmia database (MITDB), MIT-BIH supraventricular arrhythmia database (SVDB), and Charles Sturt diabetes complication screening initiative (DiSciRi) database. We proposed a simple yet effective biometric sample extraction technique for ECG samples with abnormal cardiac conditions to improve the person identification process. These sample points were then applied to four classifiers to verify the robustness of identification. Varying numbers of enrollment and recognition QRS complexes were used to validate the stability of the proposed method. Our experimentation results show that the biometric technique outperforms existing methods lacking the ability to efficiently extract features for biometric matching. This is evident by obtaining high accuracy results of 96.7% for MITDB, 96.4% for SVDB, and 99.3% for DiSciRi. Moreover, high sensitivity, specificity, positive predictive value, and Youden Indexs values further verifies the reliability of the proposed method. This technique also suggests the possibility of improving the classification performance using ECG recordings with low sampling frequency and increased number of ECG samples.

Enhancement of low sampling frequency recordings for ECG biometric matching using interpolation

Electrocardiogram (ECG) based biometric matching suffers from high misclassification error with lower sampling frequency data. This situation may lead to an unreliable and vulnerable identity authentication process in high security applications. In this paper, quality enhancement techniques for ECG data with low sampling frequency has been proposed for person identification based on piecewise cubic Hermite interpolation (PCHIP) and piecewise cubic spline interpolation (SPLINE). A total of 70 ECG recordings from 4 different public ECG databases with 2 different sampling frequencies were applied for development and performance comparison purposes. An analytical method was used for feature extraction. The ECG recordings were segmented into two parts: the enrolment and recognition datasets. Three biometric matching methods, namely, Cross Correlation (CC), Percent Root-Mean-Square Deviation (PRD) and Wavelet Distance Measurement (WDM) were used for performance evaluation before and after applying interpolation techniques. Results of the experiments suggest that biometric matching with interpolated ECG data on average achieved higher matching percentage value of up to 4% for CC, 3% for PRD and 94% for WDM. These results are compared with the existing method when using ECG recordings with lower sampling frequency. Moreover, increasing the sample size from 56 to 70 subjects improves the results of the experiment by 4% for CC, 14.6% for PRD and 0.3% for WDM. Furthermore, higher classification accuracy of up to 99.1% for PCHIP and 99.2% for SPLINE with interpolated ECG data as compared of up to 97.2% without interpolation ECG data verifies the study claim that applying interpolation techniques enhances the quality of the ECG data.

Person identification in irregular cardiac conditions using electrocardiogram signals

This paper presents a person identification mechanism in irregular cardiac conditions using ECG signals. A total of 30 subjects were used in the study from three different public ECG databases containing various abnormal heart conditions from the Paroxysmal Atrial Fibrillation Predicition Challenge database (AFPDB), MIT-BIH Supraventricular Ar-rthymia database (SVDB) and T-Wave Alternans Challenge database (TWADB). Cross correlation (CC) was used as the biometric matching algorithm with defined threshold values to evaluate the performance. In order to measure the efficiency of this simple yet effective matching algorithm, two biometric performance metrics were used which are false acceptance rate (FAR) and false reject rate (FRR). Our experimentation results suggest that ECG based biometric identification with irregular cardiac condition gives a higher recognition rate of different ECG signals when tested for three different abnormal cardiac databases yielding false acceptance rate (FAR) of 2%, 3% and 2% and false reject rate (FRR) of 1%, 2% and 0% for AFPDB, SVDB and TWADB respectively. These results also indicate the existence of salient biometric characteristics in the ECG morphology within the QRS complex that tends to differentiate individuals.

An efficient method of biometric matching using interpolated ECG data

In this paper, a person identification method using electrocardiogram (ECG) is presented based on cubic spline interpolation method. Three different databases with two different sampling rates containing 36 ECG recordings were used for development and evaluation. Each ECG recording is divided into two segments: a segment for enrolment, and a segment for recognition. The ECG features are extracted from both the training dataset and the test dataset for model development and identification. Two ECG biometric algorithms which are Cross Correlation (CC) and Percent Root-Mean-Square Deviation (PRD) were used for performance evaluation. Results of experiments confirmed that the template matching using interpolation method achieved better accuracy (up to 4.46%) than the existing method without interpolation when using ECG data with lower sampling rate.

Biometric sample extraction using Mahalanobis distance in Cardioid based graph using electrocardiogram signals

In this paper, a person identification mechanism implemented with Cardioid based graph using electrocardiogram (ECG) is presented. Cardioid based graph has given a reasonably good classification accuracy in terms of differentiating between individuals. However, the current feature extraction method using Euclidean distance could be further improved by using Mahalanobis distance measurement producing extracted coefficients which takes into account the correlations of the data set. Identification is then done by applying these extracted features to Radial Basis Function Network. A total of 30 ECG data from MITBIH Normal Sinus Rhythm database (NSRDB) and MITBIH Arrhythmia database (MITDB) were used for development and evaluation purposes. Our experimentation results suggest that the proposed feature extraction method has significantly increased the classification performance of subjects in both databases with accuracy from 97.50% to 99.80% in NSRDB and 96.50% to 99.40% in MITDB. High sensitivity, specificity and positive predictive value of 99.17%, 99.91% and 99.23% for NSRDB and 99.30%, 99.90% and 99.40% for MITDB also validates the proposed method. This result also indicates that the right feature extraction technique plays a vital role in determining the persistency of the classification accuracy for Cardioid based person identification mechanism.

Data mining technique on cardioid graph based ECG biometric authentication

In this paper, a data mining technique is used on Cardioid based person identification mechanism using electrocardiogram (ECG). Recent studies in Cardioid based ECG biometric excites a new dimension of efficient patient authentication, which places new hope in faster patient care. However, existing research suffers from lower accuracy due to random biometric template selection from fixed points in Cartesian coordinate. In this paper, we have extracted the ECG features using set of Euclidean distances with the help of data mining techniques. Euclidean distances, being independent of fixed points (as opposed to existing research) maintains higher accuracy in biometric identification when Bayes Network was implemented for classification purposes. A total of 26 ECG recordings from MIT/BIH Normal Sinus Rhythm database (NSRDB) and MIT/BIH Arrythmia database (MITDB) are used for development and evaluation. Our experimentation on these two sets of public ECG databases shows the proposed data mining based approach on Euclidean distances obtained from Cardioid graph results to 98.60% and 98.30% classification accuracy respectively.

Design and development of portable classroom attendance system based on Arduino and fingerprint biometric

In this paper, the design and development of a portable classroom attendance system based on fingerprint biometric is presented. Among the salient aims of implementing a biometric feature into a portable attendance system is security and portability. The circuit of this device is strategically constructed to have an independent source of energy to be operated, as well as its miniature design which made it more efficient in term of its portable capability. Rather than recording the attendance in writing or queuing in front of class equipped with fixed fingerprint or smart card reader. This paper introduces a portable fingerprint based biometric attendance system which addresses the weaknesses of the existing paper based attendance method or long time queuing. In addition, our biometric fingerprint based system is encrypted which preserves data integrity.

Application of Data Mining on Polynomial Based Approach for ECG Biometric

In this paper, the application of data mining techniques on polynomial based approach for better electrocardiogram (ECG) authentication mechanism is presented. Polynomials being used for ECG data processing have a history of nearly two decades. Recently it has been bringing about promising solutions for heart beat recognition problem. General polynomial based approach are used in this research and by using the polynomial coefficients extracted as unique features from the ECG signals, data mining techniques was applied for person identification. A total of 18 ECG recordings from MIT/BIH Normal Sinus Rhythm database (NSRDB) were used for development and evaluation. QRS complexes from each dataset was divided into two parts, the training and the testing dataset which was used to prove the validity of the data mining technique applied. Experimental results was classified using Multilayer Perceptron (MLP) in order to confirm the identity of an individual and was compared with the previous research using polynomials without the use of data mining technique. Our experimentation on a public ECG database suggest that the proposed data mining technique on polynomial based approach significantly improves the identification accuracy by 96% as compared to 87% from the existing study.

Acceleration plethysmogram based biometric identification

This paper presents the feasibility study of Acceleration Plethysmogram (APG) based biometric identification system. APG signals are obtained from the second derivative of the Photoplethysmogram (PPG) signal. It has been reported from previous literature that APG signals contain more information as compared to the PPG signal. Thus, in this paper, the robustness and reliability of APG signal as a biometric recognition mechanism will be proven. APG signals of 10 subjects were acquired from the Multiparameter Intelligent Monitoring in Intensive Care II Waveform Database (MIMIC2WDB) which contains PPG signals with a sampling frequency of 125 Hz. The signals were later converted into an APG waveform. Then, discriminating features are extracted from the APG morphology. Finally, these APG samples were classified using commonly known classification techniques to identify individuals. Based on the experimentation results, APG signal when using Bayes Network gives an identification rate of 97.5 percentage as compared to PPG signal of 55 percentage for the same waveform. This outcome suggests the feasibility and robustness of APG signals as a biometric modality as compared to PPG signals.