Effect of importance sampling on robust segmentation of audio-cough events in noisy environments

Monge-Álvarez, Jesús; Hoyos-Barceló, Carlos; Lesso, Paul; Escudero, Javier; Casaseca‐de‐la‐Higuera, Pablo

doi:10.1109/embc.2016.7591541

Cited by 11 publications

(16 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The baseline one used linear search (no optimization) and was used as a reference for comparison with the optimized version, and as first testbed to assess the impact of different distance metrics, normalized/non-normalized feature sets, and pruning of the database using Hart's "Condensed NN" algorithm [31]. Due to the lower performance in cough detection for k>1 [19], we based our experiments in 1-NN. The optimized versions respectively used VP-Trees, KD-Trees and Ball-Trees structures for fast search.…”

Section: A Experimental Setupmentioning

confidence: 99%

See 1 more Smart Citation

Efficient k-NN Implementation for Real-Time Detection of Cough Events in Smartphones

Hoyos-Barceló

Monge-Álvarez

Shakir

et al. 2018

IEEE J. Biomed. Health Inform.

View full text Add to dashboard Cite

The potential of telemedicine in respiratory health care has not been completely unveiled in part due to the inexistence of reliable objective measurements of symptoms such as cough. Currently available cough detectors are uncomfortable and expensive at a time when generic smartphones can perform this task. However, two major challenges preclude smartphone-based cough detectors from effective deployment namely, the need to deal with noisy environments and computational cost. This paper focuses on the latter, since complex machine learning algorithms are too slow for real-time use and kill the battery in a few hours unless specific actions are taken. In this paper, we present a robust and efficient implementation of a smartphone-based cough detector. The audio signal acquired from the device's microphone is processed by computing local Hu moments as a robust feature set in the presence of background noise. We previously demonstrated that pairing Hu moments and a standard k-NN classifier achieved accurate cough detection at the expense of computation time. To speed-up k-NN search, many tree structures have been proposed. Our cough detector uses an improved vantage point (vp)-tree with optimized construction methods and a distance function that results in faster searches. We achieve 18× speed-up over classic vp-trees, and 560× over standard implementations of k-NN in state-of-the-art machine learning libraries, with classification accuracies over 93%, enabling real-time performance on low-end smartphones.

show abstract

Section: A Experimental Setupmentioning

confidence: 99%

“…Our preliminary work in [19] proposed using local Hu moments in the time-frequency domain as a robust feature set to carry out cough detection in noisy environments. Hu-Moments have been extensively used in image processing for object recognition [20], and were recently extended for speech emotion recognition [21].…”

Section: Introductionmentioning

confidence: 99%

Efficient k-NN Implementation for Real-Time Detection of Cough Events in Smartphones

Hoyos-Barceló

Monge-Álvarez

Shakir

et al. 2018

IEEE J. Biomed. Health Inform.

View full text Add to dashboard Cite

show abstract

“…Our work in [17,18,28] adopted this approach to perform robust detection of cough events in noisy environments. However, moment computation is still a bottleneck in terms of efficiency, and this motivates our proposal to improve performance so they can be seamlessly implemented in mobile platforms.…”

Section: Image Moments For Audio Event Recognitionmentioning

confidence: 99%

Efficient computation of image moments for robust cough detection using smartphones

Hoyos-Barceló

Monge-Álvarez

Pervez

et al. 2018

Computers in Biology and Medicine

Self Cite

View full text Add to dashboard Cite

Health Monitoring apps for smartphones have the potential to improve quality of life and decrease the cost of health services. However, they have failed to live up to expectation in the context of respiratory disease. This is in part due to poor objective measurements of symptoms such as cough. Real-time cough detection using smartphones faces two main challenges namely, the necessity of dealing with noisy input signals, and the need of the algorithms to be computationally efficient, since a high battery consumption would prevent patients from using them. This paper proposes a robust and efficient smartphone-based cough detection system able to keep the phone battery consumption below 25% (16% if only the detector is considered) during 24 h use. The proposed system efficiently calculates local image moments over audio spectrograms to feed an optimized classifier for final cough detection. Our system achieves 88.94% sensitivity and 98.64% specificity in noisy environments with a 5500× speed-up and 4× battery saving compared to the baseline implementation. Power consumption is also reduced by a minimum factor of 6 compared to existing optimized systems in the literature.

show abstract

“…Similarly, Ren et al [ 18 ] use fine-grained techniques to capture the behavior of breathing during sleep through smartphones. In [ 19 ], Monge-Alvarez et al propose an automatic system for the detection of cough based on the standard audio signal of smartphones. They use a local database of sounds of coughing for comparison, and their processing uses emotion recognition algorithms.…”

Section: Related Workmentioning

confidence: 99%

Accurate Ambient Noise Assessment Using Smartphones

Zamora

Calafate

Cano

et al. 2017

Sensors

View full text Add to dashboard Cite

Nowadays, smartphones have become ubiquitous and one of the main communication resources for human beings. Their widespread adoption was due to the huge technological progress and to the development of multiple useful applications. Their characteristics have also experienced a substantial improvement as they now integrate multiple sensors able to convert the smartphone into a flexible and multi-purpose sensing unit. The combined use of multiple smartphones endowed with several types of sensors gives the possibility to monitor a certain area with fine spatial and temporal granularity, a procedure typically known as crowdsensing. In this paper, we propose using smartphones as environmental noise-sensing units. For this purpose, we focus our study on the sound capture and processing procedure, analyzing the impact of different noise calculation algorithms, as well as in determining their accuracy when compared to a professional noise measurement unit. We analyze different candidate algorithms using different types of smartphones, and we study the most adequate time period and sampling strategy to optimize the data-gathering process. In addition, we perform an experimental study comparing our approach with the results obtained using a professional device. Experimental results show that, if the smartphone application is well tuned, it is possible to measure noise levels with a accuracy degree comparable to professional devices for the entire dynamic range typically supported by microphones embedded in smartphones, i.e., 35–95 dB.

show abstract

Effect of importance sampling on robust segmentation of audio-cough events in noisy environments

Cited by 11 publications

References 15 publications

Efficient k-NN Implementation for Real-Time Detection of Cough Events in Smartphones

Efficient k-NN Implementation for Real-Time Detection of Cough Events in Smartphones

Efficient computation of image moments for robust cough detection using smartphones

Accurate Ambient Noise Assessment Using Smartphones

Contact Info

Product

Resources

About