Fahim Sufi

Novel methods of faster cardiovascular diagnosis in wireless telecardiology

With the rapid development wireless technologies, mobile phones are gaining acceptance to become an effective tool for cardiovascular monitoring. However, existing technologies have limitations in terms of efficient transmission of compressed ECG over text messaging communications like SMS and MMS. In this paper, we first propose an ECG compression algorithm which allows lossless transmission of compressed ECG over bandwidth constrained wireless link. Then, we propose several algorithms for cardiovascular abnormality detection directly from the compressed ECG maintaining end to end security, patient privacy while offering the benefits of faster diagnosis. Next, we show that our mobile phone based cardiovascular monitoring solution is capable of harnessing up to 6.72 times faster diagnosis compared to existing technologies. As the decompression time on a doctors mobile phone could be significant, our method will be highly advantageous in patient wellness monitoring system where a doctor has to read and diagnose from compressed ECGs of several patients assigned to him. Finally, we successfully implemented the prototype system by establishing mobile phone based cardiovascular patient monitoring.

ECG R-R Peak Detection on Mobile Phones

Mobile phones have become an integral part of modern life. Due to the ever increasing processing power, mobile phones are rapidly expanding its arena from a sole device of telecommunication to organizer, calculator, gaming device, web browser, music player, audio/video recording device, navigator etc. The processing power of modern mobile phones has been utilized by many innovative purposes. In this paper, we are proposing the utilization of mobile phones for monitoring and analysis of biosignal. The computation performed inside the mobile phones processor will now be exploited for healthcare delivery. We performed literature review on RR interval detection from ECG and selected few PC based algorithms. Then, three of those existing RR interval detection algorithms were programmed on JavaTM platform. Performance monitoring and comparison studies were carried out on three different mobile devices to determine their application on a realtime telemonitoring scenario.

Diagnosis of Cardiovascular Abnormalities From Compressed ECG: A Data Mining-Based Approach

Usage of compressed ECG for fast and efficient telecardiology application is crucial, as ECG signals are enormously large in size. However, conventional ECG diagnosis algorithms require the compressed ECG packets to be decompressed before diagnosis can be performed. This added step of decompression before performing diagnosis for every ECG packet introduces unnecessary delay, which is undesirable for cardiovascular diseased (CVD) patients. In this paper, we are demonstrating an innovative technique that performs real-time classification of CVD. With the help of this real-time classification of CVD, the emergency personnel or the hospital can automatically be notified via SMS/MMS/e-mail when a life-threatening cardiac abnormality of the CVD affected patient is detected. Our proposed system initially uses data mining techniques, such as attribute selection (i.e., selects only a few features from the compressed ECG) and expectation maximization (EM)-based clustering. These data mining techniques running on a hospital server generate a set of constraints for representing each of the abnormalities. Then, the patients mobile phone receives these set of constraints and employs a rule-based system that can identify each of abnormal beats in real time. Our experimentation results on 50 MIT-BIH ECG entries reveal that the proposed approach can successfully detect cardiac abnormalities (e.g., ventricular flutter/fibrillation, premature ventricular contraction, atrial fibrillation, etc.) with 97% accuracy on average. This innovative data mining technique on compressed ECG packets enables faster identification of cardiac abnormality directly from the compressed ECG, helping to build an efficient telecardiology diagnosis system.

ECG-Based Authentication

A biometric system performs template matching of acquired biometric data against template biometric data [17.1]. These biometric data can be acquired from several sources like deoxyribonucleic acid (DNA), ear, face, facial thermogram, fingerprints, gait, hand geometry, hand veins, iris, keystroke, odor, palm print, retina, signature, voice, etc. According to previous research, DNA, iris and odor provide high measurement for biometric identifiers including universalities, distinctiveness and performance [17.1]. DNA provides a one dimensional ultimate unique code for accurate identification for a person, except for the case of identical twins. In biological terms “Central Dogma” refers to the basic concept that, in nature, genetic information generally flows from the DNA to RNA (ribonucleic acid) to protein. Eventually protein is responsible for the uniqueness provided by other biometric data (finger print, iris, face, retina, etc.). Therefore, it can be inferred that the uniqueness provided by the existing biometric entities is inherited from the uniqueness of DNA. It is imperative to note that shape of the hand or palm print or face or even the shape of particular organs like the heart has distinctive features which can be useful for successful identification. The composition, mechanism and electrical activity of the human heart inherit uniqueness from the individuality of DNA. An electrocardiogram (ECG) represents the electrical activities of the heart. Figure 17.1 shows the inheritance of uniqueness for ECG inherited from the DNA.

Faster person identification using compressed ECG in time critical wireless telecardiology applications

Adoption of compression technology is often required for wireless cardiovascular monitoring, due to the enormous size of electrocardiogram (ECG) signal and limited bandwidth of Internet. However, compressed ECG must be decompressed before performing human identification using present research on ECG based biometric techniques. This additional step of decompression creates a significant processing delay for identification task. This becomes an obvious burden on a system if this needs to be done for millions of compressed ECG segments by the hospital. This paper proposes a novel method of ECG biometric directly form compressed ECG harnessing data mining (DM) techniques like attribute selection and clustering. The biometric template created by this new technique is lower in size compared to the existing ECG based biometrics as well as other forms of biometrics like face, finger, retina, etc. The template size (and also the matching time) is up to 8533 times lower than face template, 61 times lower than existing percentage root mean square (PRD) ECG based biometric template and 9 times smaller than polynomial distance measurement (PDM) based ECG biometric. Smaller template size substantially reduces the one to many matching time for biometric recognition, resulting in a faster biometric authentication mechanism and ECG stream verification directly from compressed ECG.

Polynomial distance measurement for ECG based biometric authentication

Existing electrocardiography (ECG) based biometric systems are constantly being challenged by higher misclassification error, longer acquisition time, larger template size, slower processing time and pertinence of abnormal beats within the biometric template. These challenges are the prime hindrance for ECG based biometric being commercialized as a pervasive authentication mechanism. At least,ECGbased biometric can provide a secured mechanism for cardiac patients being monitored over telephony network. In this paper, we present a polynomial distance measurement (PDM) method for ECG based biometric authentication for the very first time, according to the literature and to the best of our knowledge. The proposed PDM method is up to 12 times faster than existing algorithms, requires up to 6.5 times less template storage, needs only 2.49 (average) acquisition time with the highest accuracy rate (up to 100 per cent) when experimented on a population size of 15. Moreover, this proposed ECG based biometric system was deployed on a mobile phone based telemonitoring scenario with multilayer authentication mechanism upholding its applicability.

A Mobile Phone Based Intelligent Telemonitoring Platform

In this paper, we propose a generic smart telemonitoring platform in which the computation power of the mobile phone is highly utilized. In this approach, compression of ECG is done in real-time by the mobile phone for the very first time. The fast and effective compression scheme, designed for the proposed telemonitoring system, outperforms most of the real-time lossless ECG compression algorithms. This mobile phone based computation platform is a promising solution for privacy issues in telemonitoring through encryptions. Moreover, the mobile phones used in this platform performs preliminary detection of abnormal biosignal in realtime. Apart from the usage of mobile phones, this platform supports background biosignal abnormality surveillance using data mining agent.

Legendre Polynomials based biometric authentication using QRS complex of ECG

In this paper we propose a new Legendre Polynomials based ECG biometric technique that can efficiently be used for person indentification and authentication. we apply high-order Legendre Polynomials on QRS Complex of ECG which is considered one of the most unique signature bearing parts. We show that coefficients generated from various degrees of polynomial matchings are unique for the same person but We show that coefficients generated from various degrees of polynomial matchings are unique for the same person but are significantly different from others. We also show that even with a 4th degree ploynomial fit person authenitication/identification is possible with high degree of accuracy. This is an interesting result as we can achieve significant reduction of key sizes when coefficients generated by these fits are used as unique keys for authentication and verification of subjects.

Enforcing secured ECG transmission for realtime telemonitoring: A joint encoding, compression, encryption mechanism

Realtime telemonitoring of critical, acute and chronic patients has become increasingly popular with the emergence of portable acquisition devices and IP enabled mobile phones. During telemonitoring, enormous physiological signals are transmitted through the public communication network in realtime. However, these physiological signals can be intercepted with minimal effort, since existing telemonitoring practise ignores the privacy and security requirements. In this paper, to achieve end-to-end security, we first proposed an encoding method capable of securing Electrocardiogram (ECG) data transmission from an acquisition device to a mobile phone, and then from a mobile phone to a centralised medical server by concealing cardiovascular details as well as features in ECG data required to identify an individual. The encoding method not only conceals cardiovascular condition, but also reduces the enormous file size of the ECG with a compression ratio of up to 3.84, thus making it suitable in energy constrained small acquisition devices. As ECG data transfer faces even greater security vulnerabilities while traversing through the public Internet, we further designed and implemented 3 phase encoding-compression-encryption mechanism on mobile phones using the proposed encoding method and existing compression and encryption tools. This new mechanism elevates the security strength of the system even further. Apart from higher security, we also achieved higher compression ratio of up to 20.06, which will enable faster transmission and make the system suitable for realtime telemonitoring.

Diagnosis of cardiovascular abnormalities from compressed ECG: A data mining based approach

Usage of compressed Electrocardiography (ECG) for fast and efficient telecardiology application is crucial, as ECG signals are enormously large in size. However, conventional ECG diagnosis algorithms require the compressed ECG to be decompressed before diagnosis can be applied. This added step of decompression before performing diagnosis for every ECG packets introduces unnecessary delays, which is undesirable for cardiovascular patients. In this paper, we first used an attribute selection method that selects only a few features from the compressed ECG. Then we used clustering techniques to create normal and abnormal ECG clusters. 18 different segments (12 normal and 6 abnormal) of compressed ECG were tested with 100 % success on our model. This innovative data mining technique on compressed ECGs, now enables faster identification of cardiac abnormality directly from the compressed ECG, resulting in an efficient telecardiology diagnosis system.