Hyperglycemia Identification Using ECG in Deep Learning Era.

A growing number of smart wearable biosensors are operating in the medical IoT environment and those that capture physiological signals have received special attention. Electrocardiogram (ECG) is one of the physiological signals used in the cardiovascular and medical fields that has encouraged researchers to discover new non-invasive methods to diagnose hyperglycemia as a personal variable. Over the years, researchers have proposed different techniques to detect hyperglycemia using ECG. In this paper, we propose a novel deep learning architecture that can identify hyperglycemia using heartbeats from ECG signals. In addition, we introduce a new fiducial feature extraction technique that improves the performance of the deep learning classifier. We evaluate the proposed method with ECG data from 1119 different subjects to assess the efficiency of hyperglycemia detection of the proposed work. The result indicates that the proposed algorithm is effective in detecting hyperglycemia with a 94.53% area under the curve (AUC), 87.57% sensitivity, and 85.04% specificity. That performance represents an relative improvement of 53% versus the best model found in the literature. The high sensitivity and specificity achieved by the 10-layer deep neural network proposed in this work provide an excellent indication that ECG possesses intrinsic information that can indicate the level of blood glucose concentration.

Highly Reliable Key Generation From Electrocardiogram (ECG).

Traditional passwords are inadequate as cryptographic keys, as they are easy to forge and are vulnerable to guessing. Human biometrics have been proposed as a promising alternative due to their intrinsic nature. Electrocardiogram (ECG) is an emerging biometric that is extremely difficult to forge and circumvent, but has not yet been heavily investigated for cryptographic key generation. ECG has challenges with respect to immunity to noise, abnormalities, etc. In this paper, we propose a novel key generation approach that extracts keys from real-valued ECG features with high reliability and entropy in mind. Our technique, called interval optimized mapping bit allocation (IOMBA), is applied to normal and abnormal ECG signals under multiple session conditions. We also investigate IOMBA in the context of different feature extraction methods, such as wavelet, discrete cosine transform, etc., to find the best method for feature extraction. Experiments of IOMBA show that 217-, 38-, and 100-bit keys with 99.9%, 97.4%, and 95% average reliability and high entropy can be extracted from normal, abnormal, and multiple session ECG signals, respectively. By allowing more errors or lowering entropy, key lengths can be further increased by tunable parameters of IOMBA, which can be useful in other applications. While IOMBA is demonstrated on ECG, it should be useful for other biometrics as well.