Dynamic Q value wavelet transform for IF estimation and seizure detection via QT plane-CNN framework

Abstract The instantaneous frequency (IF) is the underlying feature for analyzing the non-stationary signals. Separating the mono components present in a non-stationary multi-component signal is a crucial step for estimating their IF. In this paper, we suggest a new decomposition method to separate...

Full description

Saved in:
Bibliographic Details
Main Authors: Amaya Rose Abraham, Anurag Nishad, Mihir Jaipuria, Abhay Upadhyay, Sudeep Baudha
Format: Article
Language:English
Published: SpringerOpen 2025-07-01
Series:EURASIP Journal on Advances in Signal Processing
Subjects:
Online Access:https://doi.org/10.1186/s13634-025-01229-4
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849767049563209728
author Amaya Rose Abraham
Anurag Nishad
Mihir Jaipuria
Abhay Upadhyay
Sudeep Baudha
author_facet Amaya Rose Abraham
Anurag Nishad
Mihir Jaipuria
Abhay Upadhyay
Sudeep Baudha
author_sort Amaya Rose Abraham
collection DOAJ
description Abstract The instantaneous frequency (IF) is the underlying feature for analyzing the non-stationary signals. Separating the mono components present in a non-stationary multi-component signal is a crucial step for estimating their IF. In this paper, we suggest a new decomposition method to separate the mono components from a multi-component non-stationary signal using a novel dynamic quality (Q) value wavelet transform (DQVWT). The signal is windowed in the time domain using a moving Gaussian function, and then, the components in the windowed signal are separated using the array of tunable-Q wavelet transform (TQWT) blocks. Here, the Q values are computed dynamically as the Gaussian window traverses over a multi-component signal. The components are added to generate mono components and then their IF is estimated. The proposed method is evaluated on multi-component non-stationary signals whose mono components are either very close in the T–F plane or overlapped in the T–F plane. Additionally, a Q vs time (QT) plane is proposed as a new feature for the classification of non-stationary signals. The efficacy of the proposed method is evaluated in terms of mean square error (MSE) and compared it with other existing methods under different signal-to-noise ratio (SNR) conditions. Through simulation results, it is shown that the MSE in IF estimation obtained using the proposed method is lower when compared with the other methods. Also, the proposed method is applied to a bat echo signal for the IF estimation, and the proposed QT contour is used for the classification of the EEG signals.
format Article
id doaj-art-3c45c63b05db4e5cbb466555ac3d5cfc
institution DOAJ
issn 1687-6180
language English
publishDate 2025-07-01
publisher SpringerOpen
record_format Article
series EURASIP Journal on Advances in Signal Processing
spelling doaj-art-3c45c63b05db4e5cbb466555ac3d5cfc2025-08-20T03:04:21ZengSpringerOpenEURASIP Journal on Advances in Signal Processing1687-61802025-07-012025112710.1186/s13634-025-01229-4Dynamic Q value wavelet transform for IF estimation and seizure detection via QT plane-CNN frameworkAmaya Rose Abraham0Anurag Nishad1Mihir Jaipuria2Abhay Upadhyay3Sudeep Baudha4Department of Electrical and Electronics Engineering, Birla Institute of Technology and Science, Pilani, K K Birla Goa CampusDepartment of Electrical and Electronics Engineering, Birla Institute of Technology and Science, Pilani, K K Birla Goa CampusDepartment of Electrical and Electronics Engineering, Birla Institute of Technology and Science, Pilani, K K Birla Goa CampusDepartment of Electronics and Communication Engineering, Indian Institute of Information TechnologyDepartment of Electrical and Electronics Engineering, Birla Institute of Technology and Science, Pilani, K K Birla Goa CampusAbstract The instantaneous frequency (IF) is the underlying feature for analyzing the non-stationary signals. Separating the mono components present in a non-stationary multi-component signal is a crucial step for estimating their IF. In this paper, we suggest a new decomposition method to separate the mono components from a multi-component non-stationary signal using a novel dynamic quality (Q) value wavelet transform (DQVWT). The signal is windowed in the time domain using a moving Gaussian function, and then, the components in the windowed signal are separated using the array of tunable-Q wavelet transform (TQWT) blocks. Here, the Q values are computed dynamically as the Gaussian window traverses over a multi-component signal. The components are added to generate mono components and then their IF is estimated. The proposed method is evaluated on multi-component non-stationary signals whose mono components are either very close in the T–F plane or overlapped in the T–F plane. Additionally, a Q vs time (QT) plane is proposed as a new feature for the classification of non-stationary signals. The efficacy of the proposed method is evaluated in terms of mean square error (MSE) and compared it with other existing methods under different signal-to-noise ratio (SNR) conditions. Through simulation results, it is shown that the MSE in IF estimation obtained using the proposed method is lower when compared with the other methods. Also, the proposed method is applied to a bat echo signal for the IF estimation, and the proposed QT contour is used for the classification of the EEG signals.https://doi.org/10.1186/s13634-025-01229-4Dynamic Q value wavelet transformNon-stationary signalsInstantaneous frequencyElectroencephalogram signalsConvolutional neural network
spellingShingle Amaya Rose Abraham
Anurag Nishad
Mihir Jaipuria
Abhay Upadhyay
Sudeep Baudha
Dynamic Q value wavelet transform for IF estimation and seizure detection via QT plane-CNN framework
EURASIP Journal on Advances in Signal Processing
Dynamic Q value wavelet transform
Non-stationary signals
Instantaneous frequency
Electroencephalogram signals
Convolutional neural network
title Dynamic Q value wavelet transform for IF estimation and seizure detection via QT plane-CNN framework
title_full Dynamic Q value wavelet transform for IF estimation and seizure detection via QT plane-CNN framework
title_fullStr Dynamic Q value wavelet transform for IF estimation and seizure detection via QT plane-CNN framework
title_full_unstemmed Dynamic Q value wavelet transform for IF estimation and seizure detection via QT plane-CNN framework
title_short Dynamic Q value wavelet transform for IF estimation and seizure detection via QT plane-CNN framework
title_sort dynamic q value wavelet transform for if estimation and seizure detection via qt plane cnn framework
topic Dynamic Q value wavelet transform
Non-stationary signals
Instantaneous frequency
Electroencephalogram signals
Convolutional neural network
url https://doi.org/10.1186/s13634-025-01229-4
work_keys_str_mv AT amayaroseabraham dynamicqvaluewavelettransformforifestimationandseizuredetectionviaqtplanecnnframework
AT anuragnishad dynamicqvaluewavelettransformforifestimationandseizuredetectionviaqtplanecnnframework
AT mihirjaipuria dynamicqvaluewavelettransformforifestimationandseizuredetectionviaqtplanecnnframework
AT abhayupadhyay dynamicqvaluewavelettransformforifestimationandseizuredetectionviaqtplanecnnframework
AT sudeepbaudha dynamicqvaluewavelettransformforifestimationandseizuredetectionviaqtplanecnnframework