ECG Biometric Authentication Using Self-Supervised Learning for IoT Edge Sensors.

Wearable Internet of Things (IoT) devices are gaining ground for continuous physiological data acquisition and health monitoring. These physiological signals can be used for security applications to achieve continuous authentication and user convenience due to passive data acquisition. This paper investigates an electrocardiogram (ECG) based biometric user authentication system using features derived from the Convolutional Neural Network (CNN) and self-supervised contrastive learning. Contrastive learning enables us to use large unlabeled datasets to train the model and establish its generalizability. We propose approaches enabling the CNN encoder to extract appropriate features that distinguish the user from other subjects. When evaluated using the PTB ECG database with 290 subjects, the proposed technique achieved an authentication accuracy of 99.15%. To test its generalizability, we applied the model to two new datasets, the MIT-BIH Arrhythmia Database and the ECG-ID Database, achieving over 98.5% accuracy without any modifications. Furthermore, we show that repeating the authentication step three times can increase accuracy to nearly 100% for both PTBDB and ECGIDDB. This paper also presents model optimizations for embedded device deployment, which makes the system more relevant to real-world scenarios. To deploy our model in IoT edge sensors, we optimized the model complexity by applying quantization and pruning. The optimized model achieves 98.67% accuracy on PTBDB, with 0.48% accuracy loss and 62.6% CPU cycles compared to the unoptimized model. An accuracy-vs-time-complexity tradeoff analysis is performed, and results are presented for different optimization levels.

Low Complexity Binarized 2D-CNN Classifier for Wearable Edge AI Devices.

Wearable Artificial Intelligence-of-Things (AIoT) devices exhibit the need to be resource and energy-efficient. In this paper, we introduced a quantized multilayer perceptron (qMLP) for converting ECG signals to binary image, which can be combined with binary convolutional neural network (bCNN) for classification. We deploy our model into a low-power and low-resource field programmable gate array (FPGA) fabric. The model requires 5.8× lesser multiply and accumulate (MAC) operations than known wearable CNN models. Our model also achieves a classification accuracy of 98.5%, sensitivity of 85.4%, specificity of 99.5%, precision of 93.3%, and F1-score of 89.2%, along with dynamic power dissipation of 34.9 µW.

ANNet: A Lightweight Neural Network for ECG Anomaly Detection in IoT Edge Sensors.

In this paper, we propose a lightweight neural network for real-time electrocardiogram (ECG) anomaly detection and system level power reduction of wearable Internet of Things (IoT) Edge sensors. The proposed network utilizes a novel hybrid architecture consisting of Long Short Term Memory (LSTM) cells and Multi-Layer Perceptrons (MLP). The LSTM block takes a sequence of coefficients representing the morphology of ECG beats while the MLP input layer is fed with features derived from instantaneous heart rate. Simultaneous training of the blocks pushes the overall network to learn distinct features complementing each other for making decisions. The network was evaluated in terms of accuracy, computational complexity, and power consumption using data from the MIT-BIH arrhythmia database. To address the class imbalance in the dataset, we augmented the dataset using SMOTE algorithm for network training. The network achieved an average classification accuracy of 97% across several records in the database. Further, the network was mapped to a fixed point model, retrained in a bit accurate fixed-point environment to compensate for the quantization error, and ported to an ARM Cortex M4 based embedded platform. In laboratory testing, the overall system was successfully demonstrated, and a significant saving of ≅ 50% power was achieved by gating the wireless transmission using the classifier. Wireless transmission was enabled only to transmit the beats deemed anomalous by the classifier. The proposed technique compares favourably with current methods in terms of computational complexity and has the advantage of stand-alone operation in the edge node, without the need for always-on wireless connectivity making it ideal for IoT wearable devices.

Study of Electrode Locations for Joint Acquisition of Impedance- and Electro-cardiography Signals.

ICG (impedance cardiography) and ECG (electrocardiography) provide important indications about functioning of the heart and of overall cardiovascular system. Measuring ICG along with ECG using wearable devices will improve the quality of health monitoring, as ICG points to important hemodynamic parameters (such as time intervals, stroke volume, cardiac output, and their variability). In this work, various electrode locations (12 different setups) have been tested for possible joint ECG & ICG data acquisition (using the same electrodes) and signal quality has been evaluated for every setup. It is shown that, while typically ICG is acquired over the whole thorax, a wrist-based joint acquisition of ECG & ICG signals can achieve acceptable signal quality and therefore can be considered in wearable sensing.

Evaluation of Level-Crossing ADCs for Event-Driven ECG Classification.

In this paper, a new methodology for choosing design parameters of level-crossing analog-to-digital converters (LC-ADCs) is presented that improves sampling accuracy and reduces the data stream rate. Using the MIT-BIH Arrhythmia dataset, several LC-ADC models are designed, simulated and then evaluated in terms of compression and signal-to-distortion ratio. A new one-dimensional convolutional neural network (1D-CNN) based classifier is presented. The 1D-CNN is used to evaluate the event-driven data from several LC-ADC models. With uniformly sampled data, the 1D-CNN has 99.49%, 92.4% and 94.78% overall accuracy, sensitivity and specificity, respectively. In comparison, a 7-bit LC-ADC with 2385 Hz clock frequency and 6-bit clock resolution offers 99.2%, 89.98% and 91.64% overall accuracy, sensitivity and specificity, respectively. It also offers 3x data compression while maintaining a signal-to-distortion ratio of 21.19 dB. Furthermore, it only requires 49% floating-point operations per second (FLOPS) for cardiac arrhythmia classification in comparison with the uniformly sampled ADC. Finally, an open-source event-driven arrhythmia database is presented.

Multimodal Multiresolution Data Fusion Using Convolutional Neural Networks for IoT Wearable Sensing.

With advances in circuit design and sensing technology, the acquisition of data from a large number of Internet of Things (IoT) sensors simultaneously to enable more accurate inferences has become mainstream. In this work, we propose a novel convolutional neural network (CNN) model for the fusion of multimodal and multiresolution data obtained from several sensors. The proposed model enables the fusion of multiresolution sensor data, without having to resort to padding/ resampling to correct for frequency resolution differences even when carrying out temporal inferences like high-resolution event detection. The performance of the proposed model is evaluated for sleep apnea event detection, by fusing three different sensor signals obtained from UCD St. Vincent University Hospital's sleep apnea database. The proposed model is generalizable and this is demonstrated by incremental performance improvements, proportional to the number of sensors used for fusion. A selective dropout technique is used to prevent overfitting of the model to any specific high-resolution input, and increase the robustness of fusion to signal corruption from any sensor source. A fusion model with electrocardiogram (ECG), Peripheral oxygen saturation signal (SpO2), and abdominal movement signal achieved an accuracy of 99.72% and a sensitivity of 98.98%. Energy per classification of the proposed fusion model was estimated to be approximately 5.61 µJ for on-chip implementation. The feasibility of pruning to reduce the complexity of the fusion models was also studied.

SomnNET: An SpO2 Based Deep Learning Network for Sleep Apnea Detection in Smartwatches.

The abnormal pause or rate reduction in breathing is known as the sleep-apnea hypopnea syndrome and affects the quality of sleep of an individual. A novel method for the detection of sleep apnea events (pause in breathing) from peripheral oxygen saturation (SpO2) signals obtained from wearable devices is discussed in this paper. The paper details an apnea detection algorithm of a very high resolution on a per-second basis for which a 1-dimensional convolutional neural network- which we termed SomnNET- is developed. This network exhibits an accuracy of 97.08% and outperforms several lower resolution state-of-the-art apnea detection methods. The feasibility of model pruning and binarization to reduce the computational complexity is explored. The pruned network with 80% sparsity exhibited an accuracy of 89.75%, and the binarized network exhibited an accuracy of 68.22%. The performance of the proposed networks is compared against several state-of-the-art algorithms.

Multistage Pruning of CNN Based ECG Classifiers for Edge Devices.

Using smart wearable devices to monitor patients' electrocardiogram (ECG) for real-time detection of arrhythmias can significantly improve healthcare outcomes. Convolutional neural network (CNN) based deep learning has been used successfully to detect anomalous beats in ECG. However, the computational complexity of existing CNN models prohibits them from being implemented in low-powered edge devices. Usually, such models are complex with lots of model parameters which results in large number of computations, memory, and power usage in edge devices. Network pruning techniques can reduce model complexity at the expense of performance in CNN models. This paper presents a novel multistage pruning technique that reduces CNN model complexity with negligible loss in performance compared to existing pruning techniques. An existing CNN model for ECG classification is used as a baseline reference. At 60% sparsity, the proposed technique achieves 97.7% accuracy and an F1 score of 93.59% for ECG classification tasks. This is an improvement of 3.3% and 9% for accuracy and F1 Score respectively, compared to traditional pruning with fine-tuning approach. Compared to the baseline model, we also achieve a 60.4% decrease in run-time complexity.

Predicting Stroke from Electronic Health Records.

Studies have identified various risk factors associated with the onset of stroke in an individual. Data mining techniques have been used to predict the occurrence of stroke based on these factors by using patients' medical records. However, there has been limited use of electronic health records to study the inter-dependency of different risk factors of stroke. In this paper, we perform an analysis of patients' electronic health records to identify the impact of risk factors on stroke prediction. We also provide benchmark performance of the state-of-art machine learning algorithms for predicting stroke using electronic health records.