Which of the three types of neural networks has the fastest converge speed?

Among the three types of neural networks, the CNN system has the fastest converge speed, whereas the LSTM system has the slowest converge speed.

How many channels are there in the CNN authentication system?

When the number of active EEG channels (Nchan) is reduced to 4, the Rank-1 accuracies of CNN, LSTM, and 1D-Convolutional LSTM authentication systems are 94.34%, 90.36%, and 94.28% respectively.

How long could it take to collect 12s EEG signals?

The proposed CNN architecture achieved 0.19% Equal Error Rate (EER), however, in practice, collecting 12s EEG signals could take too long for authentication purposes.

How many channels does the proposed LSTM authentication system have?

The proposed 1D-Convolutional LSTM authentication systems does not benefit from increasing the number of channels in the systems either, having Rank-1 accuracies of 99.58%, 99.50%, and 99.58% for 16, 32, and 64 channels respectively.

What is the trade-off between the efficiency and the performance of the authentication systems?

a trade-off can be made on increasing the efficiency and the performance of the authentication systems by increasing the number of channels in the EEG signals, or reducing the cost of the systems by using less number of electrodes of the EEG signal recording devices.

How much better is the Rank-1 accuracy of the CNN system?

The accuracy of CNN system is 0.06% higher than that of 1D-Convolutional LSTM and is 3.98% higher than that of LSTM, because the spatial information reside in the selected four channels is not sufficient for the proposed 1D-Convolutional LSTM systems to outperform CNNs.

How long does it take to load a Tensorflow model?

On the other hand, LSTM and 1D-Convolutional LSTM systems require much longer time for loading Tensorflow model, both under 20 seconds.

How many EEG channels did CNN authentication system require?

As stated in Table IV, the proposed 1D-Convolutional LSTM authentication system outperforms other previous approaches [4] by 2.58% and required only 16 EEG channels.

How many EEG recordings were identified in these experiments?

The batch size was set to 8 in these experiments, therefore, the execution time for identifying one EEG recording should be one eighth of Tbatch.

How much accuracy does the LSTM system have?

It worths mention that, comparing with CNN and 1D-Convolutional LSTM systems, the lower accuracy of the LSTM system shows that the feature extraction provided by CNNs could accelerate the learning speed for LSTM when exploiting spatial information from EEG channels.

What is the difference between the training and validation accuracy of the CNN and LSTM systems?

In addition, as shown in Fig. 4c, the training and validation losses of the 1D-Convolutional LSTM systems decrease with much less speed that those of CNN and LSTM systems, therefore requiring longer training time than CNN and LSTM systems.

Open AccessJournal Article10.1016/J.ESWA.2019.01.080

EEG-based user identification system using 1D-convolutional long short-term memory neural networks

Yingnan Sun, +2 more

- 01 Jul 2019

- Expert Systems With Applications

- Vol. 125, pp 259-267

167

TL;DR: A novel approach based on 1D Convolutional Long Short-term Memory Neural Network (1D-Convolutional LSTM) for EEG-based user identification system can greatly improve the accuracy of user identification systems by utilizing the spatiotemporal features of the EEG signals with L STM, and lowering cost of the systems by reducing the number of EEG electrodes used in the systems.

Abstract: Electroencephalographic (EEG) signals have been widely used in medical applications, yet the use of EEG signals as user identification systems for healthcare and Internet of Things (IoT) systems has only gained interests in the last few years. The advantages of EEG-based user identification systems lie in its dynamic property and uniqueness among different individuals. However, it is for this reason that manually designed features are not always adapted to the needs. Therefore, a novel approach based on 1D Convolutional Long Short-term Memory Neural Network (1D-Convolutional LSTM) for EEG-based user identification system is proposed in this paper. The performance of the proposed approach was validated with a public database consists of EEG data of 109 subjects. The experimental results showed that the proposed network has a very high averaged accuracy of 99.58%, when using only 16 channels of EEG signals, which outperforms the state-of-the-art EEG-based user identification methods. The combined use of CNNs and LSTMs in the proposed 1D-Convolutional LSTM can greatly improve the accuracy of user identification systems by utilizing the spatiotemporal features of the EEG signals with LSTM, and lowering cost of the systems by reducing the number of EEG electrodes used in the systems.

AI Agents for this Paper

Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps

Most frequently asked questions

1. What have the authors contributed in "Eeg-based authentication with 1d-convolutional long short-term memory neural networks" ?

A novel approach based on 1D Convolutional Long Short-term Memory Neural Network ( 1DConvolutional LSTM ) for EEG-based biometric identification is proposed in this paper.. The authors investigate the architecture to determine the appropriate topology to improve the efficacy of the 1D-Convolutional LSTMs for the authentication.

2. how many electrodes can be used to achieve similar performance?

In addition, with 1D-Convolutional LSTM, less number of electrodes can be used to achieve similar performance which could significantly reduce the cost of EEGbased biometric authentication systems.

3. How many epochs did the training of a network last?

The training of a network were designed to stop either it has reached 1000 epochs or the training and validation loss were no longer reduced.

4. What are the two categories of EEG-based authentication methods?

The EEG-based authentication methods can mainly be divided into two categories: manually designed features extraction approaches and deep learning approaches.

Fig. 4: Training, validation, and testing accuracy (left y-axis) and loss (right y-axis) against steps for 16-channel CNN (a), LSTM (b), and 1D-Convolutional LSTM (c) authentication systems (only the first 10× 105 steps are shown for each system))

TABLE I: Comparison of the performance of CNN, LSTM, and the proposed 1D-Convolutional LSTM authentication systems (1000 epochs) with 4, 16, 32, and 64 channels EEG signals (positions of the electrodes are shown in Fig. 3)

TABLE II: Comparison of the steps and wall time for CNN, LSTM, and the proposed 1D-Convolutional LSTM authentication systems to reach training loss (TL) less than 1, 0.1, and 0.01 (Time Unit: minute)

TABLE III: Tensorflow model loading time (Tgraph) and averaged execution time for batch testing (Tbatch) for CNN, LSTM, and the proposed 1D-Convolutional LSTM authentication systems (Time Unit: second)

Fig. 1: Overview of the proposed EEG-based biometric authentication system

TABLE IV: Comparison with some state-of-the-art EEG-based authentication systems

Citations

•Journal Article•10.1109/TCBB.2021.3052811

EEG-Based Brain-Computer Interfaces (BCIs): A Survey of Recent Studies on Signal Sensing Technologies and Computational Intelligence Approaches and Their Applications

Xiaotong Gu, +6 more

- 01 Sep 2021

- IEEE/ACM Transactions on Computational B...

TL;DR: The current status of BCI and signal sensing technologies for collecting reliable EEG signals are reviewed and state-of-the-art computational intelligence techniques, including fuzzy models and transfer learning in machine learning and deep learning algorithms, are demonstrated to detect, monitor, and maintain human cognitive states and task performance in prevalent applications.

...read moreread less

309

Journal Article•10.1016/J.ESWA.2019.113114

A comprehensive survey on the biometric recognition systems based on physiological and behavioral modalities

Shaveta Dargan, +1 more

- 01 Apr 2020

- Expert Systems With Applications

TL;DR: A comprehensive and deep survey that compactly and systematically summarizes the literature work done on unimodal and multimodal biometric systems and analyzes the feature extraction techniques, classifiers, datasets, results, efficiency and reliability of the system with high and multi-dimensional perspectives is explicated.

...read moreread less

247

•Journal Article•10.1016/J.CHAOS.2020.110227

Time Series Forecasting of Covid-19 using Deep Learning Models: India-USA Comparative Case Study

Sourabh Shastri, +4 more

- 20 Aug 2020

- Chaos Solitons & Fractals

TL;DR: Convolution LSTM outperformed the other two models and predicts the Covid-19 cases with high accuracy and very less error for all four datasets of both countries.

...read moreread less

242

•Journal Article•10.1109/ACCESS.2019.2960617

Security and Privacy for the Internet of Medical Things Enabled Healthcare Systems: A Survey

Yingnan Sun, +2 more

- 18 Dec 2019

- IEEE Access

TL;DR: The security and privacy challenges, requirements, threats, and future research directions in the domain of IoMT are reviewed providing a general overview of the state-of-the-art approaches.

...read moreread less

238

•Journal Article•10.3390/V12070769

Deploying Machine and Deep Learning Models for Efficient Data-Augmented Detection of COVID-19 Infections

Ahmed Sedik, +11 more

- 16 Jul 2020

- Viruses

TL;DR: The proposed algorithm is effective in performing a rapid and consistent Corona virus diagnosis that is primarily aimed at assisting clinicians in making accurate identification of the virus.

...read moreread less

203

...

Expand

References

•Proceedings Article

Adam: A Method for Stochastic Optimization

Diederik P. Kingma, +1 more

- 01 Jan 2015

TL;DR: This work introduces Adam, an algorithm for first-order gradient-based optimization of stochastic objective functions, based on adaptive estimates of lower-order moments, and provides a regret bound on the convergence rate that is comparable to the best known results under the online convex optimization framework.

...read moreread less

138.5K

•Posted Content

Adam: A Method for Stochastic Optimization

Diederik P. Kingma, +1 more

- 22 Dec 2014

- arXiv: Learning

TL;DR: In this article, the adaptive estimates of lower-order moments are used for first-order gradient-based optimization of stochastic objective functions, based on adaptive estimate of lowerorder moments.

...read moreread less

82.5K

•Journal Article•10.1161/01.CIR.101.23.E215

PhysioBank, PhysioToolkit, and PhysioNet: components of a new research resource for complex physiologic signals.

Ary L. Goldberger, +9 more

- 13 Jun 2000

- Circulation

TL;DR: The newly inaugurated Research Resource for Complex Physiologic Signals (RRSPS) as mentioned in this paper was created under the auspices of the National Center for Research Resources (NCR Resources).

...read moreread less

14.3K

Journal Article•10.1109/TBME.2004.827072

BCI2000: a general-purpose brain-computer interface (BCI) system

Gerwin Schalk, +4 more

- 24 May 2004

- IEEE Transactions on Biomedical Engineer...

TL;DR: This report is intended to describe to investigators, biomedical engineers, and computer scientists the concepts that the BCI2000 system is based upon and gives examples of successful BCI implementations using this system.

...read moreread less

3.1K

Proceedings Article•10.1109/ACV.1994.341300

Parameterisation of a stochastic model for human face identification

F.S. Samaria, +1 more

- 05 Dec 1994

TL;DR: This paper presents a set of experimental results in which various HMM parameterisations are analysed and shows that stochastic modelling can be used successfully to encode feature information.

...read moreread less

...

Expand

EEG-based user identification system using 1D-convolutional long short-term memory neural networks

Chat with Paper

AI Agents for this Paper

Most frequently asked questions

1. What have the authors contributed in "Eeg-based authentication with 1d-convolutional long short-term memory neural networks" ?

2. how many electrodes can be used to achieve similar performance?

3. How many epochs did the training of a network last?

4. What are the two categories of EEG-based authentication methods?

5. Which of the three types of neural networks has the fastest converge speed?

6. How many channels are there in the CNN authentication system?

7. How long could it take to collect 12s EEG signals?

8. How many channels does the proposed LSTM authentication system have?

9. What is the trade-off between the efficiency and the performance of the authentication systems?

10. How much better is the Rank-1 accuracy of the CNN system?

11. How long does it take to load a Tensorflow model?

12. How many EEG channels did CNN authentication system require?

13. How many EEG recordings were identified in these experiments?

14. How much accuracy does the LSTM system have?

15. What is the difference between the training and validation accuracy of the CNN and LSTM systems?

Figures

Citations

EEG-Based Brain-Computer Interfaces (BCIs): A Survey of Recent Studies on Signal Sensing Technologies and Computational Intelligence Approaches and Their Applications

A comprehensive survey on the biometric recognition systems based on physiological and behavioral modalities

Time Series Forecasting of Covid-19 using Deep Learning Models: India-USA Comparative Case Study

Security and Privacy for the Internet of Medical Things Enabled Healthcare Systems: A Survey

Deploying Machine and Deep Learning Models for Efficient Data-Augmented Detection of COVID-19 Infections

References

Adam: A Method for Stochastic Optimization

Adam: A Method for Stochastic Optimization

PhysioBank, PhysioToolkit, and PhysioNet: components of a new research resource for complex physiologic signals.

BCI2000: a general-purpose brain-computer interface (BCI) system

Parameterisation of a stochastic model for human face identification

Related Papers (5)

PhysioBank, PhysioToolkit, and PhysioNet: components of a new research resource for complex physiologic signals.

BCI2000: a general-purpose brain-computer interface (BCI) system

An EEG-Based Biometric System Using Eigenvector Centrality in Resting State Brain Networks

Affective EEG-Based Person Identification Using the Deep Learning Approach

Deep learning with convolutional neural networks for EEG decoding and visualization.