RESEARCH COMMUNICATIONS CURRENT SCIENCE, VOL. 112, NO. 9, 10 MAY 2017 1915 *For correspondence. (e-mail: chinmay_sscet@yahoo.co.in) System design approach for heartbeat detection and classification of individuals irrespective of their physical condition Chinmay Chandrakar* and Monisha Sharma Department of Electronics and Telecommunication, Shri Shankaracharaya College of Engineering and Technology, Chhattisgarh Swami Vivekananda Technical University, Bhilai 490 020, India In an electrocardiogram (ECG), the heartbeat feature QRS is an important parameter for analysis in any heartbeat classification automated diagnosis system. In this communication the method which we have proposed is by using the counter which is used in par- allel. The first level is detection of heartbeats, which uses hashing of ECG features. In the second level, the profiler profiles a person’s regular and irregular ECG characteristic behaviour. The proposed method works on data related with ECG, instead of particular fea- tures of ECG. Because of parallel processing, data storage unit requirements and the processing time are less. The dependent values in the proposed method vary according to the changes in the ECG waveform. Such type of analysis is suitable for detection of heart disease. The most significant application of such char- acteristic plotting is to generate an alert signal for ir- regular ECG behaviour in a person. Such automated system will be useful in remote areas where a cardi- ologist may not be easily available. Keywords: Data storage units, electrocardiogram signal, parallel processing, QRS detection. THE determination and separation of QRS waveform be- tween regular and irregular waveforms are particularly important clinical criteria for patient diagnosis. Over the last few decades, several techniques have been proposed to determine these waveforms 1–3 . For instance, Senhadji et al. 4 compared the performance of three different wave- lets to determine the beats which are hidden in the wave- form. In this paper, we propose a method for ECG waveform separation by using derivative of lower order in which Gaussian function was used 5 . Many more meth- ods have been proposed which concentrated on spectral 6,7 or wavelet features 8,9 , related with superimposed fea- tures 10,11 and spatial context 12,13 and to differentiate the heartbeats signal 14–19 . Pan and Tompkins 14 first proposed the method of determination of heartbeat in real time in processor. Ning and Selesnick 20 proposed that the loca- tion of the true peak can be determined which has the largest magnitude within its 200 ms time window. The ECG feature extractor provided by LabVIEW Biomedical toolkit detects QRS waves 21 . It was reported that for QRS determination, valley points before and after R-wave are sufficient 22 . It was also reported that the preprocessed ECG signal is converted into a train of pulses using the IF sampler 23 . Once the process of beat determination is complete, the correct beat discrimination process of heartbeats takes place. Methods such as detection of any symbol, repeated pattern which is going to repeat, network based on neurons, and vector machines have been sug- gested for heartbeat discrimination 24–26 . Learning meth- ods based on training data from the sample which result in new mapping based on information which is processed has been discussed earlier 27 . de Chazal et al. 28 discrimi- nated the beats by studying the distance between the two R-peaks, ECG waveform. Christov et al. 29 proposed com- parison of ECG features for beat classification in time and frequency domain. Haseena et al. 30 discussed a com- bination of neuron and fuzzy for discrimination of heart- beats. We have used arrhythmia database of the Massachu- setts Institute of Technology–Beth Israel Hospital (MIT– BIH) to test the efficiency of the proposed technique 3 . All ECG data used here are sampled at 360 Hz, and the reso- lution of each is 8 bits/sample, therefore the bit rate of these data is 2880 bps. The method was run for both regular and irregular heartbeats. High-frequency noise was removed using the simple time constant equation in which the window is moved from one R-peak to the next, while the low-frequency noise was removed using the frequency-domain transfor- mation 31 . The proposed method depends more on a series of data of ECG waveform rather than any specific feature of the ECG waveform. The basic concept is to partition the input ECG signals into series of 0–1 strings. The next step is to select a string of L bytes. The start and end lim- its of this binary information are selected so that when long binary strings are repeated, they should lie on the same start and end limits. For this, the start and end limits are selected based on R-peak. Between these two repeated start and end points, the minimum value of binary data bits Q and S points starting from the middle R-peak to the left and right are determined for repeated occurrence within a fixed time duration. Hereafter, the data bits between Q and S points will be called as signature which is W bytes. L is varying as it depends on the start and end points of the R-peak, which basically works on R–R interval. Similarly, W is not fixed because it depends on the location of Q and S points. For the process to work in real-time QRS detection and classification, we used the storage elements which are arranged in parallel. There- fore in order to avoid the repeated number of counting in a memory elements string was hashed to a certain value. Two steps of hashing have been done in two phases. Phase 1 hash was first generated by the string. This hash