Abstract
The work presented in this paper aims to show the effectiveness of dimensionality reduction in convolutional neural network (CNN) based vowel classification from covert/imagined speech. Imagined speech is referred to as phonological classes, words or sentences pronounced internally. It is acquired in a non-invasive manner by placing Electroencephalogram (EEG) sensors over the cerebral cortex region of the head. Covert speech is a spontaneous imagination or active thoughts of speaking of a human being without any articulatory movements. Therefore, identifying phonological classification attracted many applications for those who have the inability to speak due to lock-in syndrome or motor muscular impairments. The present study develops a CNN-based vowel classification system by processing EEG representing the imagined speech. In the proposed methodology, the CNN features extracted from the spectrograms (Time-Frequency) representation of each EEG channel data have been subjected to dimensionality reduction using principal component analysis (PCA). Dimensionally reduced CNN features are further subjected to linear discriminant analysis for transformation and classification. A significant improvement in the imaginary vowel classification performance is confirmed for linear discriminant analysis (LDA) based classification of dimensionality reduced CNN feature vectors. The CNN-based vowel classification performance is just above the chance level. The variational mode decomposition (VMD) based preprocessing of the EEG channels prior to classification further improved the performance of vowel recognition. The performances are observed to be consistent on all 15 subjects in the open access Coretto DB EEG database where each imagined utterance has been acquired using 6 EEG Channels.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Similar content being viewed by others
References
Biswas, S., Sinha, R.: Lateralization of brain during EEG based covert speech classification. In: Proceedings of 15th IEEE India Council International Conference (INDICON), pp. 1–5 (2018)
Biswas, S., Sinha, R.: Wavelet filterbank-based EEG rhythm-specific spatial features for covert speech classification. IET Sig. Process. 16(1), 92–105 (2022)
Cooney, C., Folli, R., Coyle, D.: Optimizing layers improves CNN generalization and transfer learning for imagined speech decoding from EEG. In: Proceedings of the IEEE International Conference on Systems, Man and Cybernetics (SMC), pp. 1311–1316 (2019)
Coretto, G.A.P., Gareis, I.E., Rufiner, H.L.: Open access database of EEG signals recorded during imagined speech. In: Proceedings of the 12th International Symposium on Medical Information Processing and Analysis, vol. 10160, p. 1016002 (2017)
Dragomiretskiy, K., Zosso, D.: Variational mode decomposition. IEEE Trans. Sig. Process. 62(3), 531–544 (2014)
Duda, R.O., Hart, P.E., Stork, D.G.: Pattern Classification. wiley, New York (2001)
Govind, D., Pravena, D., Ajay, S.G.: Improved epoch extraction using variational mode decomposition based spectral smoothing of zero frequency filtered emotive speech signals. In: Proceedings of the National Conference on Communications (NCC) (2018)
James, G., Witten, D., Hastie, T., Tibshirani, R.: An introduction to statistical learning with applications in R. Springer (2014)
Lawhern, V.J., Solon, A.J., Waytowich, N.R., Gordon, S.M., Hung, C.P., Lance, B.J.: Eegnet: a compact convolutional neural network for EEG-based brain-computer interfaces. J. Neural Eng. 15(5),(2018)
LeCun, Y., Huang, F.J., Bottou, L.: Learning methods for generic object recognition with invariance topose and lighting. In: Proceedings of the IEEE International Conference on Computer Vision and Pattern Recognition (2004)
LeCun, Y., Kavukcuoglu, K., Farabet, C.: Convolutional neural networks and applications in vision. In: Proceedings of the IEEE International Symposium on Circuits and Systems (ISCAS) (2010)
Lee, D.Y., Lee, M., Lee, S.W.: Decoding imagined speech based on deep metric learning for intuitive BCI communication. IEEE Trans. Neural Syst. Rehab. Eng. 29, 1363–1374 (2021)
Lee, D.Y., Lee, M., Lee, S.W.: Decoding imagined speech based on deep metric learning for intuitive. BCI Commun. 29, 1363–1374 (2021)
Nicolas-Alonso, L.F., Gomez-Gil, J.: Brain computer interfaces, a review. Sensors. 12(2), 1211–1279 (2012)
Panachakel, J.T., Ramakrishnan, A.: Decoding covert speech from EEG-a comprehensive review. Front. NeuroSci. (2021). https://doi.org/10.3389/fnins.2021.642251
Panachakel, J.T., Ramakrishnan, A., Ananthapadmanabha, T.: Decoding imagined speech using wavelet features and deep neural networks. In: Proceedings of the IEEE 16th India Council International Conference (INDICON), pp. 1–4 (2019)
Pankaj, D., Govind, D., Narayanankutty, K.A.: A novel method for removing RICIAN noise from MRI based on variational mode decomposition. Biomed. Sig. Process. Control 69,(2021)
Rabiner, L.R., Juang, B.H.: Fundamentals of Speech Recognition. PTR Prentice Hall,Englewood Cliffs, N.J. (1993)
Zhao, S., Rudzicz, F.: Classifying phonological categories in imagined and articulated speech. In: Proceedings of the International Conference on Acoustics, Speech and Signal Processing (ICASSP) (2015)
Acknowledgments
The presente work is funded by the ongoing MEITY, Govt. of India, funded NLTM-BHASHINI consortium project titled “Speech technologies in Indian languages: speech quality control”. The authors were motivated to work in the decoding of imagined speech from the Global Initiative for Academic Networks (GIAN) course on Cognitive Speech Processing conducted by Prof. H. L. Rufiner, Research Institute for Signals, Systems and Computational Intelligence, National University of Litoral (UNL), Santa Fe, Argentina and Prof. S. R Mahadeva Prasanna. Department of Electrical Engineering, IIT Dharwad, INDIA. GIAN course was organized in April 2022.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2022 Springer Nature Switzerland AG
About this paper
Cite this paper
Banerjee, O., Govind, D., Dubey, A.K., Gangashetty, S.V. (2022). Significance of Dimensionality Reduction in CNN-Based Vowel Classification from Imagined Speech Using Electroencephalogram Signals. In: Prasanna, S.R.M., Karpov, A., Samudravijaya, K., Agrawal, S.S. (eds) Speech and Computer. SPECOM 2022. Lecture Notes in Computer Science(), vol 13721. Springer, Cham. https://doi.org/10.1007/978-3-031-20980-2_5
Download citation
DOI: https://doi.org/10.1007/978-3-031-20980-2_5
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-20979-6
Online ISBN: 978-3-031-20980-2
eBook Packages: Computer ScienceComputer Science (R0)