Studying critical frequency bands and channels for EEG-based automobile sound recognition with machine learning

2022 ◽  
Vol 185 ◽  
pp. 108389
Author(s):  
Liping Xie ◽  
Chihua Lu ◽  
Zhien Liu ◽  
Lirong Yan ◽  
Tao Xu
Keyword(s):  
Machine Learning ◽  
Critical Frequency ◽  
Sound Recognition ◽  
Frequency Bands

scholarly journals Environmental Sound Recognition on Embedded Systems: From FPGAs to TPUs

Electronics ◽  
2021 ◽  
Vol 10 (21) ◽  
pp. 2622
Author(s):  
Jurgen Vandendriessche ◽  
Nick Wouters ◽  
Bruno da Silva ◽  
Mimoun Lamrini ◽  
Mohamed Yassin Chkouri ◽  
...  
Keyword(s):  
Machine Learning ◽  
High Performance ◽  
Machine Learning Techniques ◽  
Sound Recognition ◽  
Learning Approaches ◽  
Environmental Sound ◽  
Embedded Devices ◽  
Power Efficient ◽  
Computationally Intensive ◽  
Environmental Sound Recognition

In recent years, Environmental Sound Recognition (ESR) has become a relevant capability for urban monitoring applications. The techniques for automated sound recognition often rely on machine learning approaches, which have increased in complexity in order to achieve higher accuracy. Nonetheless, such machine learning techniques often have to be deployed on resource and power-constrained embedded devices, which has become a challenge with the adoption of deep learning approaches based on Convolutional Neural Networks (CNNs). Field-Programmable Gate Arrays (FPGAs) are power efficient and highly suitable for computationally intensive algorithms like CNNs. By fully exploiting their parallel nature, they have the potential to accelerate the inference time as compared to other embedded devices. Similarly, dedicated architectures to accelerate Artificial Intelligence (AI) such as Tensor Processing Units (TPUs) promise to deliver high accuracy while achieving high performance. In this work, we evaluate existing tool flows to deploy CNN models on FPGAs as well as on TPU platforms. We propose and adjust several CNN-based sound classifiers to be embedded on such hardware accelerators. The results demonstrate the maturity of the existing tools and how FPGAs can be exploited to outperform TPUs.

Intra-domain and cross-domain transfer learning for time series

2021 ◽  
Author(s):  
Erik Otović ◽  
Marko Njirjak ◽  
Dario Jozinović ◽  
Goran Mauša ◽  
Alberto Michelini ◽  
...  
Keyword(s):  
Machine Learning ◽  
Time Series ◽  
Transfer Learning ◽  
Time Series Data ◽  
The Other ◽  
Series Data ◽  
Sound Recognition ◽  
Transfer Of Knowledge ◽  
Learning Models ◽  
Machine Learning Models

<p>In this study, we compared the performance of machine learning models trained using transfer learning and those that were trained from scratch - on time series data. Four machine learning models were used for the experiment. Two models were taken from the field of seismology, and the other two are general-purpose models for working with time series data. The accuracy of selected models was systematically observed and analyzed when switching within the same domain of application (seismology), as well as between mutually different domains of application (seismology, speech, medicine, finance). In seismology, we used two databases of local earthquakes (one in counts, and the other with the instrument response removed) and a database of global earthquakes for predicting earthquake magnitude; other datasets targeted classifying spoken words (speech), predicting stock prices (finance) and classifying muscle movement from EMG signals (medicine).<br>In practice, it is very demanding and sometimes impossible to collect datasets of tagged data large enough to successfully train a machine learning model. Therefore, in our experiment, we use reduced data sets of 1,500 and 9,000 data instances to mimic such conditions. Using the same scaled-down datasets, we trained two sets of machine learning models: those that used transfer learning for training and those that were trained from scratch. We compared the performances between pairs of models in order to draw conclusions about the utility of transfer learning. In order to confirm the validity of the obtained results, we repeated the experiments several times and applied statistical tests to confirm the significance of the results. The study shows when, within the set experimental framework, the transfer of knowledge brought improvements in terms of model accuracy and in terms of model convergence rate.<br><br>Our results show that it is possible to achieve better performance and faster convergence by transferring knowledge from the domain of global earthquakes to the domain of local earthquakes; sometimes also vice versa. However, improvements in seismology can sometimes also be achieved by transferring knowledge from medical and audio domains. The results show that the transfer of knowledge between other domains brought even more significant improvements, compared to those within the field of seismology. For example, it has been shown that models in the field of sound recognition have achieved much better performance compared to classical models and that the domain of sound recognition is very compatible with knowledge from other domains. We came to similar conclusions for the domains of medicine and finance. Ultimately, the paper offers suggestions when transfer learning is useful, and the explanations offered can provide a good starting point for knowledge transfer using time series data.</p>

scholarly journals Identification of selection and inhibition components in a Go/NoGo task from EEG spectra using a machine learning model

2019 ◽  
Author(s):  
Bambi L. DeLaRosa ◽  
Jeffrey S. Spence ◽  
Michael A. Motes ◽  
Wing To ◽  
Sven Vanneste ◽  
...  
Keyword(s):  
Machine Learning ◽  
Neural Mechanisms ◽  
Frequency Bands ◽  
Time Frequency ◽  
Eeg Spectra ◽  
Therapeutic Uses ◽  
Machine Learning Model ◽  
Superior Parietal Cortex ◽  
Task Conditions ◽  
Traditional Approaches

AbstractPrior Go/NoGo studies have localized specific regions and EEG spectra for which traditional approaches have distinguished between Go and NoGo conditions. A more detailed characterization of the spatial distribution and timing of the synchronization of frequency bands would contribute substantially to the clarification of neural mechanisms that underlie performance of the Go/NoGo task. The present study used a machine learning approach to learn the features that distinguish between ERSPs involved in selection and inhibition in a Go/NoGo task. A neural network classifier was used to predict task conditions for each subject to characterize ERSPs associated with Go versus NoGo trials. The final model accurately identified individual task conditions at an overall rate of 92%, estimated by 5-fold cross-validation. The detailed accounting of EEG time-frequency patterns localized to brain sources (i.e., thalamus, preSMA, orbitofrontal cortex, and superior parietal cortex) provides elaboration on previous findings from fMRI and EEG studies and more information about EEG power changes in multiple frequency bands (i.e., primarily theta power increase, alpha decreases, and beta increases and decreases) within these regions underlying the selection and inhibition processes engaged in the Go and NoGo trials. This extends previous findings, providing more information about neural mechanisms underlying selection and inhibition processes engaged in the Go and NoGo trials, respectively, and may offer insight into therapeutic uses of neuromodulation in neural dysfunction.

scholarly journals MosAIc: A Classical Machine Learning Multi-Classifier Based Approach against Deep Learning Classifiers for Embedded Sound Classification

2021 ◽  
Vol 11 (18) ◽  
pp. 8394
Author(s):  
Lancelot Lhoest ◽  
Mimoun Lamrini ◽  
Jurgen Vandendriessche ◽  
Nick Wouters ◽  
Bruno da Silva ◽  
...  
Keyword(s):  
Machine Learning ◽  
Deep Learning ◽  
Sound Recognition ◽  
Learning Models ◽  
Environmental Sound ◽  
Machine Learning Classifiers ◽  
Sound Classification ◽  
Learning Classifiers ◽  
Constrained Devices ◽  
Environmental Sound Recognition

Environmental Sound Recognition has become a relevant application for smart cities. Such an application, however, demands the use of trained machine learning classifiers in order to categorize a limited set of audio categories. Although classical machine learning solutions have been proposed in the past, most of the latest solutions that have been proposed toward automated and accurate sound classification are based on a deep learning approach. Deep learning models tend to be large, which can be problematic when considering that sound classifiers often have to be embedded in resource constrained devices. In this paper, a classical machine learning based classifier called MosAIc, and a lighter Convolutional Neural Network model for environmental sound recognition, are proposed to directly compete in terms of accuracy with the latest deep learning solutions. Both approaches are evaluated in an embedded system in order to identify the key parameters when placing such applications on constrained devices. The experimental results show that classical machine learning classifiers can be combined to achieve similar results to deep learning models, and even outperform them in accuracy. The cost, however, is a larger classification time.

scholarly journals Acoustic Cues to Beat Induction: A Machine Learning Perspective

2006 ◽  
Vol 24 (2) ◽  
pp. 177-188 ◽  
Author(s):  
Fabien Gouyon ◽  
Gerhard Widmer ◽  
Xavier Serra ◽  
Arthur Flexer
Keyword(s):  
Machine Learning ◽  
Common Knowledge ◽  
Acoustic Properties ◽  
Acoustic Cues ◽  
Acoustic Features ◽  
Frequency Bands ◽  
Feature Sets ◽  
Special Relevance ◽  
Beat Induction ◽  
Large Corpus

This article brings forward the question of which acoustic features are the most adequate for identifying beats computationally in acoustic music pieces. We consider many different features computed on consecutive short portions of acoustic signal, among which those currently promoted in the literature on beat induction from acoustic signals and several original features, unmentioned in this literature. Evaluation of feature sets regarding their ability to provide reliable cues to the localization of beats is based on a machine learning methodology with a large corpus of beat-annotated music pieces, in audio format, covering distinctive music categories. Confirming common knowledge, energy is shown to be a very relevant cue to beat induction (especially the temporal variation of energy in various frequency bands, with the special relevance of frequency bands below 500 Hz and above 5 kHz). Some of the new features proposed in this paper are shown to outperform features currently promoted in the literature on beat induction from acoustic signals.We finally hypothesize that modeling beat induction may involve many different, complementary acoustic features and that the process of selecting relevant features should partly depend on acoustic properties of the very signal under consideration.

scholarly journals Machine Learning Framework for Sensing and Modeling Interference in IoT Frequency Bands

2020 ◽  
pp. 1-1 ◽  
Author(s):  
Bassel Al Homssi ◽  
Akram Al-Hourani ◽  
Zarko Krusevac ◽  
Wayne S T Rowe
Keyword(s):  
Machine Learning ◽  
Frequency Bands ◽  
Learning Framework

scholarly journals Design of a Low-Cost Configurable Acoustic Sensor for the Rapid Development of Sound Recognition Applications

Electronics ◽  
2020 ◽  
Vol 9 (7) ◽  
pp. 1155 ◽  
Author(s):  
Rosa Maria Alsina-Pagès ◽  
Marcos Hervás ◽  
Leticia Duboc ◽  
Jordi Carbassa
Keyword(s):  
Machine Learning ◽  
Low Cost ◽  
Rapid Development ◽  
Learning Algorithms ◽  
Noise Pollution ◽  
Urban Environments ◽  
Machine Learning Algorithms ◽  
Sound Recognition ◽  
Acoustic Sensor ◽  
Acoustic Sensor Networks

Concerned about the noise pollution in urban environments, the European Commission (EC) has created an Environmental Noise Directive 2002/49/EC (END) requiring Member states to publish noise maps and noise management plans every five years for cities with a high density of inhabitants, major roads, railways and airports. The END also requires the noise pressure levels for these sources to be presented independently. Currently, data measurements and the representations of the noise pressure levels in such maps are performed semi-manually by experts. This process is time and cost consuming, as well as limited to presenting only a static picture of the noise levels. To overcome these issues, we propose the deployment of Wireless Acoustic Sensor Networks with several nodes in urban environments that can enable the generation of real-time noise level maps, as well as detect the source of the sound thanks to machine learning algorithms. In this paper, we briefly review the state of the art of the hardware used in wireless acoustic applications and propose a low-cost sensor based on an ARM cortex-A microprocessor. This node is able to process machine learning algorithms for sound source detection in-situ, allowing the deployment of highly scalable sound identification systems.

Sign in / Sign up

Export Citation Format

Share Document