Acoustic Biometric System Based on Preprocessing Techniques and Linear Support Vector Machines

Val, Lara del; Fuente, Alberto Izquierdo; Villacorta, Juan J.; Raboso, Mariano

doi:10.3390/s150614241

Cited by 12 publications

(6 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The CNN and LSTM classifiers are trained and tested using beamformed spectrogram images of acoustic signal detections as inputs, whereas logistic regression, SVM, and decision tree classifiers are trained and tested using 12 features extracted from each acoustic signal detection in beamformed spectrograms. Note that the SVM classifier can also be trained using images [53][54][55] as input, such as beamformed spectrogram images, and will be investigated in future work.…”

Section: Training and Test Data Setmentioning

confidence: 99%

Comparing Performances of Five Distinct Automatic Classifiers for Fin Whale Vocalizations in Beamformed Spectrograms of Coherent Hydrophone Array

et al. 2020

View full text Add to dashboard Cite

A large variety of sound sources in the ocean, including biological, geophysical, and man-made, can be simultaneously monitored over instantaneous continental-shelf scale regions via the passive ocean acoustic waveguide remote sensing (POAWRS) technique by employing a large-aperture densely-populated coherent hydrophone array system. Millions of acoustic signals received on the POAWRS system per day can make it challenging to identify individual sound sources. An automated classification system is necessary to enable sound sources to be recognized. Here, the objectives are to (i) gather a large training and test data set of fin whale vocalization and other acoustic signal detections; (ii) build multiple fin whale vocalization classifiers, including a logistic regression, support vector machine (SVM), decision tree, convolutional neural network (CNN), and long short-term memory (LSTM) network; (iii) evaluate and compare performance of these classifiers using multiple metrics including accuracy, precision, recall and F1-score; and (iv) integrate one of the classifiers into the existing POAWRS array and signal processing software. The findings presented here will (1) provide an automatic classifier for near real-time fin whale vocalization detection and recognition, useful in marine mammal monitoring applications; and (2) lay the foundation for building an automatic classifier applied for near real-time detection and recognition of a wide variety of biological, geophysical, and man-made sound sources typically detected by the POAWRS system in the ocean.

show abstract

Section: Training and Test Data Setmentioning

confidence: 99%

Comparing Performances of Five Distinct Automatic Classifiers for Fin Whale Vocalizations in Beamformed Spectrograms of Coherent Hydrophone Array

et al. 2020

View full text Add to dashboard Cite

show abstract

“…It is observed that if the frequency increases, the spatial resolution improves and the main parts of the body could be discerned. The use of parameter extraction algorithms will improve the classification, as shown in the authors’ previous work, using 1D microphone arrays [12]. …”

Section: Case Study: Biometric Identification Of Peoplementioning

confidence: 99%

“…The authors of this paper have experience in the design [6] and development of acoustic imaging systems, based in acoustic arrays, used in many different fields, such as detection and tracking systems [7,8], Ambient Assisted Living [9], or biometric identification systems [10,11,12]. The arrays used were ULA (Uniform Linear Array), formed by acoustic sensors distributed uniformly along a line.…”

Section: Introductionmentioning

confidence: 99%

Design and Evaluation of a Scalable and Reconfigurable Multi-Platform System for Acoustic Imaging

Izquierdo

Villacorta

Val

et al. 2016

Sensors

Self Cite

View full text Add to dashboard Cite

This paper proposes a scalable and multi-platform framework for signal acquisition and processing, which allows for the generation of acoustic images using planar arrays of MEMS (Micro-Electro-Mechanical Systems) microphones with low development and deployment costs. Acoustic characterization of MEMS sensors was performed, and the beam pattern of a module, based on an 8 × 8 planar array and of several clusters of modules, was obtained. A flexible framework, formed by an FPGA, an embedded processor, a computer desktop, and a graphic processing unit, was defined. The processing times of the algorithms used to obtain the acoustic images, including signal processing and wideband beamforming via FFT, were evaluated in each subsystem of the framework. Based on this analysis, three frameworks are proposed, defined by the specific subsystems used and the algorithms shared. Finally, a set of acoustic images obtained from sound reflected from a person are presented as a case study in the field of biometric identification. These results reveal the feasibility of the proposed system.

show abstract

“…The authors of this paper have experience in the design and development of acoustic ULAs (Uniform Linear Arrays) [4,5,6,7,8]. These arrays are simple, but they are limited to estimate the spatial position of the sound source in only one dimension (azimuth or elevation).…”

Section: Introductionmentioning

confidence: 99%

Implementation of a Virtual Microphone Array to Obtain High Resolution Acoustic Images

Izquierdo

Villacorta

Val

et al. 2017

Sensors

Self Cite

View full text Add to dashboard Cite

Using arrays with digital MEMS (Micro-Electro-Mechanical System) microphones and FPGA-based (Field Programmable Gate Array) acquisition/processing systems allows building systems with hundreds of sensors at a reduced cost. The problem arises when systems with thousands of sensors are needed. This work analyzes the implementation and performance of a virtual array with 6400 (80 × 80) MEMS microphones. This virtual array is implemented by changing the position of a physical array of 64 (8 × 8) microphones in a grid with 10 × 10 positions, using a 2D positioning system. This virtual array obtains an array spatial aperture of 1 × 1 m2. Based on the SODAR (SOund Detection And Ranging) principle, the measured beampattern and the focusing capacity of the virtual array have been analyzed, since beamforming algorithms assume to be working with spherical waves, due to the large dimensions of the array in comparison with the distance between the target (a mannequin) and the array. Finally, the acoustic images of the mannequin, obtained for different frequency and range values, have been obtained, showing high angular resolutions and the possibility to identify different parts of the body of the mannequin.

show abstract

Acoustic Biometric System Based on Preprocessing Techniques and Linear Support Vector Machines

Cited by 12 publications

References 22 publications

Comparing Performances of Five Distinct Automatic Classifiers for Fin Whale Vocalizations in Beamformed Spectrograms of Coherent Hydrophone Array

Comparing Performances of Five Distinct Automatic Classifiers for Fin Whale Vocalizations in Beamformed Spectrograms of Coherent Hydrophone Array

Design and Evaluation of a Scalable and Reconfigurable Multi-Platform System for Acoustic Imaging

Implementation of a Virtual Microphone Array to Obtain High Resolution Acoustic Images

Contact Info

Product

Resources

About