EDITH :ECG biometrics aided by Deep learning for reliable Individual auTHentication

Ibtehaz, Nabil; Chowdhury, Muhammad E. H.; Khandakar, Amith; Kiranyaz, Serkan; Rahman, M. Sohel; Tahir, Anas; Qiblawey, Yazan; Rahman, Tawsifur

doi:10.48550/arxiv.2102.08026

Cited by 3 publications

(3 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For convolutional neural networks, we can use various methods like saliency maps [24] or score-CAM [25] methods to study model interpretation. However, for multi-layer perceptrons, it is non-trivial to do so.…”

Section: Model Interpretationmentioning

confidence: 99%

RamanNet: A generalized neural network architecture for Raman Spectrum Analysis

Ibtehaz¹,

Chowdhury²,

Khandakar³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Raman spectroscopy provides a vibrational profile of the molecules and thus can be used to uniquely identify different kind of materials. This sort of fingerprinting molecules has thus led to widespread application of Raman spectrum in various fields like medical dignostics, forensics, mineralogy, bacteriology and virology etc. Despite the recent rise in Raman spectra data volume, there has not been any significant effort in developing generalized machine learning methods for Raman spectra analysis. We examine, experiment and evaluate existing methods and conjecture that neither current sequential models nor traditional machine learning models are satisfactorily sufficient to analyze Raman spectra. Both has their perks and pitfalls, therefore we attempt to mix the best of both worlds and propose a novel network architecture RamanNet. RamanNet is immune to invariance property in CNN and at the same time better than traditional machine learning models for the inclusion of sparse connectivity. Our experiments on 4 public datasets demonstrate superior performance over the much complex state-of-the-art methods and thus RamanNet has the potential to become the defacto standard in Raman spectra data analysis.

show abstract

Section: Model Interpretationmentioning

confidence: 99%

RamanNet: A generalized neural network architecture for Raman Spectrum Analysis

Ibtehaz¹,

Chowdhury²,

Khandakar³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Similarly, Donida et al utilized a convolutional neural network (CNN) for biometric authentication and binarized the ECG signal to speed up the matching process [12]. Moreover, the authentication schemes [13][14][15] also utilized CNN as a classification method after extracting features from ECG signals. Hammad et al [16] proposed a multimodal biometric scheme that fused both ECG and Fingerprints data to apply the CNN feature extraction method.…”

Section: Introductionmentioning

confidence: 99%

Advanced Authentication Scheme with Bio-Key Using Artificial Neural Network

et al. 2022

View full text Add to dashboard Cite

The improvements in the field of health monitoring have revolutionized our daily lifestyle by developing various applications that did not exist before. However, these applications have serious security concerns; they also can be taken good care of by utilizing the Electrocardiogram (ECG) as potential biometrics. The ECG provides robustness against forgery attacks unlike conventional methods of authentication. Therefore, it has attained the utmost attention and is utilized in several authentication solutions. In this paper, we have presented an efficient architecture for an advanced authentication scheme that utilized a binarized form (bio-key) of ECG signal along with an Artificial Neural Network (ANN) to enhance the authentication process. In order to prove the concept, we have developed the testbed and acquired ECG signals using the AD8232 ECG recording module under a controlled environment. The variable-length bio-keys are extracted using an algorithm after the feature extraction process. The extracted features along with bio-keys are utilized for template formation and also for training/testing of the ANN model to enhance the accuracy of the authentication process. The performance of authentication results depicted high authentication accuracy of 98% and minimized the equal error rate (EER) to 2%. Moreover, our scheme outperformed comparative peers’ work in terms of accuracy and EER.

show abstract

“…Bashar et al, [24] explored the combination of brain signals (EEG) and heart signals (ECG) for authentication using wavelet domain statistical features with multiple classifiers and achieved a 90.5% F1 score. Ibtehaz et al [25] proposed a framework called EDITH, which performs competitively using just a single ECG heartbeat (96-99.75% accuracy) and can be further enhanced by fusing multiple beats (100% accuracy from 3 to 6 beats.…”

Section: Introductionmentioning

confidence: 99%

Multimodal EEG and Keystroke Dynamics Based Biometric System Using Machine Learning Algorithms

et al. 2021

Self Cite

View full text Add to dashboard Cite

Electroencephalography (EEG) based biometric systems are gaining attention for their anti-spoofing capability but lack accuracy due to signal variability at different psychological and physiological conditions. On the other hand, keystroke dynamics-based systems achieve very high accuracy but have low anti-spoofing capability. To address these issues, a novel multimodal biometric system combining EEG and keystroke dynamics is proposed in this paper. A dataset was created by acquiring both keystroke dynamics and EEG signals simultaneously from 10 users. Each user participated in 500 trials at 10 different sessions (days) to replicate real-life signal variability. A machine learning classification pipeline is developed using multi-domain feature extraction (time, frequency, time-frequency), feature selection (Gini impurity), classifier design, and score level fusion. Different classifiers were trained, validated, and tested for two different classification experimentspersonalized and generalized. For identification and authentication, 99.9% and 99.6% accuracies are achieved, respectively for Random Forest classifier using 5 fold cross-validation. These results outperform the individual modalities with a significant margin (~5%). We also developed a binary template matching-based algorithm, which gives 6X faster system with 93.64% accuracy. The proposed method can be considered secure and reliable for any kind of biometric identification and authentication.

show abstract

EDITH :ECG biometrics aided by Deep learning for reliable Individual auTHentication

Cited by 3 publications

References 38 publications

RamanNet: A generalized neural network architecture for Raman Spectrum Analysis

RamanNet: A generalized neural network architecture for Raman Spectrum Analysis

Advanced Authentication Scheme with Bio-Key Using Artificial Neural Network

Multimodal EEG and Keystroke Dynamics Based Biometric System Using Machine Learning Algorithms

Contact Info

Product

Resources

About