Sound Environment Analysis in Smart Home

Sehili, Mohamed El Amine; Lecouteux, Benjamin; Vacher, Michel; Portet, François; Istrate, Dan; Dorizzi, Bernadette; Boudy, Jérôme

doi:10.1007/978-3-642-34898-3_14

Cited by 28 publications

(18 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Once the sound event is classified as speech (respectively, nonspeech sound), the sound object is sent to the ASR module to extract voice commands (respectively, to the sound classification module to determine the class of sound). Due to the focus of the study on the speech interaction, the sound classification will not be detailed in this article (see Sehili et al [2012] for details about the sound classification) while the ASR module is detailed in chapter 3.2.…”

Section: The Audio Processing System: Patshmentioning

confidence: 99%

Evaluation of a Context-Aware Voice Interface for Ambient Assisted Living

Vacher

Caffiau

Portet

et al. 2015

ACM Trans. Access. Comput.

Self Cite

View full text Add to dashboard Cite

This article presents an experiment with seniors and people with visual impairment in a voice-controlled smart home using the SWEET-HOME system. The experiment shows some weaknesses in automatic speech recognition that must be addressed, as well as the need for better adaptation to the user and the environment. Users were disturbed by the rigid structure of the grammar and were eager to adapt it to their own preferences. Surprisingly, while no humanoid aspect was introduced in the system, the senior participants were inclined to embody the system. Despite these aspects to improve, the system has been favorably assessed as diminishing most participant fears related to the loss of autonomy. . 2015. Evaluation of a context-aware voice interface for Ambient Assisted Living: qualitative user study vs. quantitative system evaluation.

show abstract

Section: The Audio Processing System: Patshmentioning

confidence: 99%

Evaluation of a Context-Aware Voice Interface for Ambient Assisted Living

Vacher

Caffiau

Portet

et al. 2015

ACM Trans. Access. Comput.

Self Cite

View full text Add to dashboard Cite

show abstract

“…On the on hand, most of the sound-based recognizers are limited in recognizing human speech, together with some of its characteristics like speaker recognition and his emotional state in order to obtain detailed information about their subject of interest. On the other hand, there are very few studies, which aim to examine in an abstract way the daily human activities in a home environment according to their acoustic characteristics [9,10,12,16,18,20,21]. Despite their generalized way of analyzing sounds, they are all developed in a healthcare perspective and often make the implication that certain sound implies certain activity, which is not necessarily true.…”

Section: Sound-based Activity Recognitionmentioning

confidence: 98%

Analyzing Sounds of Home Environment for Device Recognition

Dimitrov¹,

Britz²,

Brandherm³

et al. 2014

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. Home environments are one of the subjects of study regarding ambient intelligent systems for various purposes, including development of assistance systems for the elderly and energy consumption optimization. Sensing the environmental state via different sensors is the first and crucial component of every ambient intelligent system. In this work we investigate the use of environmental sounds for touch-free audio-based device recognition in a home environment. For this purpose, we analyzed sound characteristics of typical home appliances using different processing techniques. We are using the acquired knowledge to develop a flexible set of features, which can be set manually or determined automatically. To classify the device-specific acoustic fingerprintsconsisting of a significant subset of our features -we use established supervised learning techniques, whereby we optimized the straightforward ones. After building a recognition basis for the recognition of fixed length sound buffers on demand, we implemented a live recognition mode for real-time environment monitoring, providing runtime setup adjustments. We then extended our work with the recognition of untrained, simultaneously working, known devices by mixing their records, utilizing semi-supervised learning. We then anticipated promising results in our evaluation in various aspects, including recognition rate, performance for the different combinations of features, as well as to study the reliability of an automatic mixing of trained data.

show abstract

“…Audio-based activity recognition has received a lot of attention from researchers in recent years [16,17]. A number of studies have also taken the first steps to characterize the indoor sound environment and the classification of events [18,19].…”

Section: Related Workmentioning

confidence: 99%

Audio content analysis for unobtrusive event detection in smart homes

Vafeiadis

Votis

Giakoumis

et al. 2020

Engineering Applications of Artificial Intelligence

View full text Add to dashboard Cite

Environmental sound signals are multi-source, heterogeneous, and varying in time. Many systems have been proposed to process such signals for event detection in ambient assisted living applications. Typically, these systems use feature extraction, selection, and classification. However, despite major advances, several important questions remain unanswered, especially in real-world settings. This paper contributes to the body of knowledge in the field by addressing the following problems for ambient sounds recorded in various real-world kitchen environments: (1) which features and which classifiers are most suitable in the presence of background noise? (2) what is the effect of signal duration on recognition accuracy? (3) how do the signal-to-noise-ratio and the distance between the microphone and the audio source affect the recognition accuracy in an environment in which the system was not trained? We show that for systems that use traditional classifiers, it is beneficial to combine gammatone frequency cepstral coefficients and discrete wavelet transform coefficients and to use a gradient boosting classifier. For systems based on deep learning, we consider 1D and 2D Convolutional Neural Networks (CNN) using mel-spectrogram energies and mel-spectrograms images as inputs, respectively, and show that the 2D CNN outperforms the 1D CNN. We obtained competitive classification results

show abstract

Sound Environment Analysis in Smart Home

Cited by 28 publications

References 21 publications

Evaluation of a Context-Aware Voice Interface for Ambient Assisted Living

Evaluation of a Context-Aware Voice Interface for Ambient Assisted Living

Analyzing Sounds of Home Environment for Device Recognition

Audio content analysis for unobtrusive event detection in smart homes

Contact Info

Product

Resources

About