A Machine Hearing System for Robust Cough Detection Based on a High-Level Representation of Band-Specific Audio Features.

Cough is a protective reflex conveying information on the state of the respiratory system. Cough assessment has been limited so far to subjective measurement tools or uncomfortable (i.e., non-wearable) cough monitors. This limits the potential of real-time cough monitoring to improve respiratory care. OBJECTIVE: This paper presents a machine hearing system for audio-based robust cough segmentation that can be easily deployed in mobile scenarios. METHODS: Cough detection is performed in two steps. First, a short-term spectral feature set is separately computed in five predefined frequency bands: [0, 0.5), [0.5, 1), [1, 1.5), [1.5, 2), and [2, 5.5125] kHz. Feature selection and combination are then applied to make the short-term feature set robust enough in different noisy scenarios. Second, high-level data representation is achieved by computing the mean and standard deviation of short-term descriptors in 300 ms long-term frames. Finally, cough detection is carried out using a support vector machine trained with data from different noisy scenarios. The system is evaluated using a patient signal database which emulates three real-life scenarios in terms of noise content. RESULTS: The system achieves 92.71% sensitivity, 88.58% specificity, and 90.69% Area Under Receiver Operating Charcteristic (ROC) curve (AUC), outperforming state-of-the-art methods. CONCLUSION: Our research outcome paves the way to create a device for cough monitoring in real-life situations. SIGNIFICANCE: Our proposal is aligned with a more comfortable and less disruptive patient monitoring, with benefits for patients (allows self-monitoring of cough symptoms), practitioners (e.g., assessment of treatments or better clinical understanding of cough patterns), and national health systems (by reducing hospitalizations).

Robust Detection of Audio-Cough Events Using Local Hu Moments.

Telehealth has shown potential to improve access to healthcare cost-effectively in respiratory illness. However, it has failed to live up to expectation, in part because of poor objective measures of symptoms such as cough events, which could lead to early diagnosis or prevention. Considering the burden that these conditions constitute for national health systems, an effort is needed to foster telehealth potential by developing low-cost technology for efficient monitoring and analysis of cough events. This paper proposes the use of local Hu moments as a robust feature set for automatic cough detection in smartphone-acquired audio signals. The final system feeds a k-nearest-neighbor classifier with the extracted features. To properly evaluate the system in a diversity of noisy backgrounds, we contaminated real cough audio data with a variety of sounds including noise from both indoor and outdoor environments and noncough events (sneeze, laugh, speech, etc.). The created database allows flexible settings of signal-to-noise ratio levels between background sounds and events (cough and noncough). This evaluation was complemented using real patient data from an outpatient clinic. The system is able to detect cough events with high sensitivity (up to 88.51%) and specificity (up to 99.77%) in a variety of noisy environments, overcoming other state-of-the-art audio features. Our proposal paves the way for ubiquitous cough monitoring with minimal disruption in daily activities.

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

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.

Efficient computation of image moments for robust cough detection using smartphones.

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.

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

This paper proposes a new cough detection system based on audio signals acquired from conventional smartphones. The system relies on local Hu moments to characterize cough events and a Λ-NN classifier to distinguish cough events from non-cough ones (speech, laugh, sneeze, etc.) and noisy sounds. To deal with the unbalance between classes, we employ Distinct-Borderline2 Synthetic Minority Oversampling Technique and a bespoke cost matrix. The system additionally features a post-processing module to avoid isolated false negatives and, this way, increases sensitivity. Evaluation has been carried out using a database comprising a variety of cough and non-cough events and different types of background noise. In this study, we specifically focused on noise likely to appear when the user is carrying the smartphone in daily activities. Different Signal to Noise Ratio values were tested ranging between -15 and 0 dB. Our experiments confirm that local Hu moments are suitable not only for characterizing cough events but also for coping with noisy environments. Results show a sensitivity of 94.17% and a specificity of 92.16% at -15 dB. Thus, our system shows potential as a reliable and place-ubiquitous monitoring device that helps patients self-manage their own respiratory diseases and avoids unreported or fabricated symptoms.

Effect of downsampling and compressive sensing on audio-based continuous cough monitoring.

This paper presents an efficient cough detection system based on simple decision-tree classification of spectral features from a smartphone audio signal. Preliminary evaluation on voluntary coughs shows that the system can achieve 98% sensitivity and 97.13% specificity when the audio signal is sampled at full rate. With this baseline system, we study possible efficiency optimisations by evaluating the effect of downsampling below the Nyquist rate and how the system performance at low sampling frequencies can be improved by incorporating compressive sensing reconstruction schemes. Our results show that undersampling down to 400 Hz can still keep sensitivity and specificity values above 90% despite of aliasing. Furthermore, the sparsity of cough signals in the time domain allows keeping performance figures close to 90% when sampling at 100 Hz using compressive sensing schemes.