Enabling Real-Time Computation of Psycho-Acoustic Parameters in Acoustic Sensors Using Convolutional Neural Networks

Lopez-Ballester, Jesus; Pastor-Aparicio, Adolfo; Felici-Castell, Santiago; Segura-García, Jaume; Cobos, Máximo

doi:10.1109/jsen.2020.2995779

Cited by 20 publications

(15 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Ref. [ 59 ] significantly improves the speed of calculation of acoustic parameters in Zwicker’s psycho-acoustic nuisance model on IoT devices such as Raspberry Pi [ 60 ]. It is very challenging to accurately calculate soundscape profiling in real time within a traditional WASN (wireless acoustic sensor networks) environment.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Time-Constrained Adversarial Defense in IoT Edge Devices through Kernel Tensor Decomposition and Multi-DNN Scheduling

Kim

Joo

2022

Sensors

View full text Add to dashboard Cite

The development of deep learning technology has resulted in great contributions in many artificial intelligence services, but adversarial attack techniques on deep learning models are also becoming more diverse and sophisticated. IoT edge devices take cloud-independent on-device DNN (deep neural network) processing technology to exhibit a fast response time. However, if the computational complexity of the denoizer for adversarial noises is high, or if a single embedded GPU is shared by multiple DNN models, adversarial defense at the on-device level is bound to represent a long latency. To solve this problem, eDenoizer is proposed in this paper. First, it applies Tucker decomposition to reduce the computational amount required for convolutional kernel tensors in the denoizer. Second, eDenoizer effectively orchestrates both the denoizer and the model defended by the denoizer simultaneously. In addition, the priority of the CPU side can be projected onto the GPU which is completely priority-agnostic, so that the delay can be minimized when the denoizer and the defense target model are assigned a high priority. As a result of confirming through extensive experiments, the reduction of classification accuracy was very marginal, up to 1.78%, and the inference speed accompanied by adversarial defense was improved up to 51.72%.

show abstract

Section: Related Workmentioning

confidence: 99%

“…It is very challenging to accurately calculate soundscape profiling in real time within a traditional WASN (wireless acoustic sensor networks) environment. The authors of [ 59 ] used an end-to-end CNN-based solution to allow calculation of four PA (psycho-acoustic annoyance) parameters 250 times faster than conventional algorithms.…”

Section: Related Workmentioning

confidence: 99%

Time-Constrained Adversarial Defense in IoT Edge Devices through Kernel Tensor Decomposition and Multi-DNN Scheduling

Kim

Joo

2022

Sensors

View full text Add to dashboard Cite

show abstract

“…The afore mentioned parameters are calculated based on the impulse response of the room with the only exception of the SII [20], which is extracted from a speech signal recorded in the position of the room where we wish to analyze intelligibility. The calculation of parameters based on impulse response is much more complex than SII, so it is not possible to do it in a simple, fast or effective way in a node of our WASN, as it happens for example with the parameters of psychoacoustic disturbance [21]. In fact, we proposed to implement a convolutional neuronal network (CNN) that would allow us to successfully predict them, just as we did with the parameters of psychoacoustic disturbance [21].…”

Section: Introductionmentioning

confidence: 99%

“…The calculation of parameters based on impulse response is much more complex than SII, so it is not possible to do it in a simple, fast or effective way in a node of our WASN, as it happens for example with the parameters of psychoacoustic disturbance [21]. In fact, we proposed to implement a convolutional neuronal network (CNN) that would allow us to successfully predict them, just as we did with the parameters of psychoacoustic disturbance [21]. Psychoacoustic annoyance quantifies how disturbing different sounds can be to humans.…”

Section: Introductionmentioning

confidence: 99%

Speech Intelligibility Analysis and Approximation to Room Parameters through the Internet of Things

et al. 2021

Self Cite

View full text Add to dashboard Cite

In recent years, Wireless Acoustic Sensor Networks (WASN) have been widely applied to different acoustic fields in outdoor and indoor environments. Most of these applications are oriented to locate or identify sources and measure specific features of the environment involved. In this paper, we study the application of a WASN for room acoustic measurements. To evaluate the acoustic characteristics, a set of Raspberry Pi 3 (RPi) has been used. One is used to play different acoustic signals and four are used to record at different points in the room simultaneously. The signals are sent wirelessly to a computer connected to a server, where using MATLAB we calculate both the impulse response (IR), and different acoustic parameters, such as the Speech Intelligibility Index (SII). In this way, the evaluation of room acoustic parameters with asynchronous IR measurements two different applications has been explored. Finally, the network features have been evaluated to assess the effectiveness of this system.

show abstract

“…• WASN can be a good tool to make large-scale and long-term measurements [4], but some devices can have limitation when computing complex calculation, such as advanced acoustic parameters. However, these computing limitations can be overcome with on-edge computing, machine learning algorithms and improved features of future devices [45].…”

mentioning

confidence: 99%

Long-Term Temporal Analysis of Psychoacoustic Parameters of the Acoustic Environment in a University Campus Using a Wireless Acoustic Sensor Network

Montoya-Belmonte

Navarro

2020

Sustainability

View full text Add to dashboard Cite

Higher education institution campuses have been growing in size to become small cities that encompass all the infrastructures and offer many services to the university community. In order to achieve sustainable growth, it is necessary to take into account different environmental aspects, including the acoustic environment. In this paper, the application of a wireless acoustic sensor network is proposed for the long-term analysis of psychoacoustic parameters of the acoustic environment in a university campus. A network of acoustic nodes is designed and four nodes are deployed at both indoor and outdoor locations of the campus for environmental noise monitoring. A measurement campaign has been running continuously for nine months, including three months in which the state of emergency due to COVID-19 pandemic was established and the university switched to online classes. Acquired and stored data include not only the basic sound level parameters, e.g., equivalent or percentile sound pressure levels, but also psychoacoustic parameters, that are more related with noise assessment of the environment, e.g., loudness or sharpness sound levels. After analysing the acoustic environment of the campus through the complete set of nodes, results show two zones with higher noise indicators where some action plan should be developed. A detailed analysis of the temporal evolution of noise levels in these acoustically saturated zones is done, comparing values between a period of regular activity and the period of state of emergency. Different frequency distributions of sound levels are detected between day and evening periods providing a better insight into the noise sources. The results also show that binaural psychoacoustic annoyance values are correlated with loudness values in these locations, however sharpness values have minor effects.

show abstract

Enabling Real-Time Computation of Psycho-Acoustic Parameters in Acoustic Sensors Using Convolutional Neural Networks

Cited by 20 publications

References 17 publications

Time-Constrained Adversarial Defense in IoT Edge Devices through Kernel Tensor Decomposition and Multi-DNN Scheduling

Time-Constrained Adversarial Defense in IoT Edge Devices through Kernel Tensor Decomposition and Multi-DNN Scheduling

Speech Intelligibility Analysis and Approximation to Room Parameters through the Internet of Things

Long-Term Temporal Analysis of Psychoacoustic Parameters of the Acoustic Environment in a University Campus Using a Wireless Acoustic Sensor Network

Contact Info

Product

Resources

About